CN118134652A

CN118134652A - Asset configuration scheme generation method and device, electronic equipment and medium

Info

Publication number: CN118134652A
Application number: CN202410296672.5A
Authority: CN
Inventors: 陈芃宇; 游屹; 唐鹏飞; 袁超
Original assignee: China Construction Bank Corp; CCB Finetech Co Ltd
Current assignee: China Construction Bank Corp; CCB Finetech Co Ltd
Priority date: 2024-03-15
Filing date: 2024-03-15
Publication date: 2024-06-04

Abstract

The application relates to the technical field of data analysis, in particular to a method, a device, electronic equipment and a medium for generating an asset configuration scheme, which are used for solving the problem that the quality of the asset configuration scheme generated in the related technology is unstable and cannot be matched with a customer highly. The method comprises the following steps: basic information and asset information of a target object are acquired; based on the basic information and the asset information, carrying out industry category preference prediction on the target object, and determining industry category preference information of the target object, wherein the industry category preference information characterizes the preference degree of the target object to the industry category to which the financial product belongs; selecting at least one target financial product from a plurality of financial products included in the target financial product type according to the industry category preference information; an asset allocation scheme is output based on the at least one target financial product, the asset allocation scheme indicating a specific gravity of each target financial product in the assets of the target object.

Description

Asset configuration scheme generation method and device, electronic equipment and medium

Technical Field

The present application relates to the field of data analysis technologies, and in particular, to a method and apparatus for generating an asset allocation scheme, an electronic device, and a medium.

Background

With the continuous development of the financial market and the continuous improvement of the living standard of people, the purchasing demands of people for financial products are also increasing. Especially for high equity customers, it is important how to distribute their assets reasonably due to the large amount of own assets.

When generating the asset configuration scheme of the customer at present, a series of questionnaires are issued to the customer to obtain relevant information of the customer, subjective analysis is performed by a customer manager, and the existing financial products are matched with the customer, so that a final asset configuration scheme is obtained and pushed to the customer.

However, when generating asset allocation schemes in this manner, it takes a lot of time for the customer to fill out the questionnaire, and the generated asset allocation schemes are also closely related to the experience and analysis capabilities of the customer manager, which, once they have failed due to lack of experience or capabilities, may provide the customer with asset allocation schemes of lower quality or that do not match the customer.

Therefore, how to provide a high-quality asset allocation scheme with high matching degree to customers is needed to be solved.

Disclosure of Invention

The embodiment of the application provides a method, a device, equipment and a medium for generating an asset configuration scheme, which are used for solving the problem that the quality of the asset configuration scheme generated in the related technology is unstable and cannot be matched with a client in a high degree.

In a first aspect, an embodiment of the present application provides a method for generating an asset configuration scheme, where the method includes:

basic information and asset information of a target object are acquired;

Based on the basic information and the asset information, carrying out industry category preference prediction on the target object, and determining industry category preference information of the target object, wherein the industry category preference information characterizes the preference degree of the target object on the industry category to which the financial product belongs;

Selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information;

An asset allocation scheme is output based on the at least one target financial product, the asset allocation scheme indicating a specific gravity of each target financial product in the assets of the target object.

In a possible implementation manner, before the selecting at least one target financial product from the plurality of financial products included in the target financial product type according to the industry category preference information, the method further includes: selecting a target financial product type for the target object based on the risk preference of the target object; the target risk preference characterizes a risk acceptance of the target object; the selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information includes: and selecting a target financial product belonging to the target industry category from a plurality of financial products included in the target financial product type according to the target industry category recorded in the industry category preference information.

In a possible implementation, the target risk preference of the target object is determined according to the following method: extracting features of the basic information and the asset information, and determining a feature set of the target object; analyzing the feature set of the target object based on the association relation between the feature set of the different objects and the risk preferences of the different objects to obtain the risk preferences of the target object.

In a possible implementation manner, the analyzing the feature set of the target object to obtain the risk preference of the target object based on the association relationship between the feature set of the different objects and the risk preference of the different objects includes: performing risk preference classification on the target object according to the feature set through a trained risk preference classification model, and determining target risk preference of the target object; the risk preference classification model is trained based on feature sets and risk preference labels of training samples, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

In a possible implementation manner, the performing industry category preference prediction on the target object based on the basic information and the asset information, and determining the industry category preference information of the target object includes: extracting features of the basic information and the asset information, and determining a feature set of the target object; and analyzing the characteristic set of the target object based on the association relation between the characteristic set of the different objects and the industry category preference of the different objects to obtain the industry category preference information of the target object.

In a possible implementation manner, the analyzing the feature set of the target object to obtain the industry category preference information of the target object based on the association relation between the feature set of the different objects and the industry category preference of the different objects includes: carrying out industry category preference prediction on the target object according to the feature set by using a trained industry category preference prediction model, and determining industry category preference information of the target object; the industry class preference prediction model is trained based on a feature set of training samples and an industry class preference label, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

In a possible implementation manner, the feature extracting the basic information and the asset information to determine a feature set of the target object includes: extracting features of the basic information and the asset information to obtain a candidate feature set of the target object; screening one or more features from the candidate feature set to form the feature set; wherein, the characteristics in the characteristic set satisfy: the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set or the one or more features are associated with an industry category preference that is greater than the other features in the candidate feature set or the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set and the industry category preference that is greater than the other features in the candidate feature set.

In a possible implementation manner, the screening one or more features from the candidate feature set to form the feature set includes: based on each candidate feature included in the candidate feature set, respectively generating random features, and taking the candidate features with influence degree on model variables being greater than that of the random features as undetermined features; constructing a classification model through an automatic modeling model, and sequencing importance degrees of all undetermined features based on the influence degree of the undetermined features on the model output accuracy; the classification model is a model for classifying industry category preferences and/or risk preferences; based on the ranking result, a plurality of features included in the feature set are determined.

In a second aspect, an embodiment of the present application provides an asset allocation scheme generating apparatus, the apparatus including:

The communication unit is used for acquiring basic information and asset information of the target object;

The preference information determining unit is used for predicting industry category preference of the target object based on the basic information and the asset information, determining industry category preference information of the target object, and the industry category preference information characterizes preference degree of the target object to an industry category to which the financial product belongs;

An asset allocation unit for selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information; an asset allocation scheme is output based on the at least one target financial product, the asset allocation scheme indicating a specific gravity of each target financial product in the assets of the target object.

In a possible implementation manner, before the asset configuration unit selects at least one target financial product from the plurality of financial products included in the target financial product type according to the industry category preference information, the asset configuration unit is further configured to: selecting a target financial product type for the target object based on the risk preference of the target object; the target risk preference characterizes a risk acceptance of the target object; the asset allocation unit is configured to, when selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information: and selecting a target financial product belonging to the target industry category from a plurality of financial products included in the target financial product type according to the target industry category recorded in the industry category preference information.

In a possible implementation manner, the target risk preference of the target object is determined by the preference information determining unit according to the following method: extracting features of the basic information and the asset information, and determining a feature set of the target object; analyzing the feature set of the target object based on the association relation between the feature set of the different objects and the risk preferences of the different objects to obtain the risk preferences of the target object.

In a possible implementation manner, the preference information determining unit is configured to, when analyzing the feature set of the target object to obtain the risk preference of the target object based on an association relationship between the feature set of the different object and the risk preference of the different object: performing risk preference classification on the target object according to the feature set through a trained risk preference classification model, and determining target risk preference of the target object; the risk preference classification model is trained based on feature sets and risk preference labels of training samples, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

In a possible implementation manner, the preference information determining unit performs industry category preference prediction on the target object based on the basic information and the asset information, and when determining the industry category preference information of the target object, is used for: extracting features of the basic information and the asset information, and determining a feature set of the target object; and analyzing the characteristic set of the target object based on the association relation between the characteristic set of the different objects and the industry category preference of the different objects to obtain the industry category preference information of the target object.

In a possible implementation manner, the preference information determining unit is configured to, when analyzing the feature set of the target object to obtain the industry category preference information of the target object, based on an association relationship between the feature set of the different object and the industry category preference of the different object: carrying out industry category preference prediction on the target object according to the feature set by using a trained industry category preference prediction model, and determining industry category preference information of the target object; the industry class preference prediction model is trained based on a feature set of training samples and an industry class preference label, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

In a possible implementation manner, the preference information determining unit performs feature extraction on the basic information and the asset information, and when determining a feature set of the target object, is configured to: extracting features of the basic information and the asset information to obtain a candidate feature set of the target object; screening one or more features from the candidate feature set to form the feature set; wherein, the characteristics in the characteristic set satisfy: the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set or the one or more features are associated with an industry category preference that is greater than the other features in the candidate feature set or the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set and the industry category preference that is greater than the other features in the candidate feature set.

In a possible implementation manner, the asset configuration unit is configured to screen one or more features from the candidate feature set, and when forming the feature set, is configured to: based on each candidate feature included in the candidate feature set, respectively generating random features, and taking the candidate features with influence degree on model variables being greater than that of the random features as undetermined features; constructing a classification model through an automatic modeling model, and sequencing importance degrees of all undetermined features based on the influence degree of the undetermined features on the model output accuracy; the classification model is a model for classifying industry category preferences and/or risk preferences; based on the ranking result, a plurality of features included in the feature set are determined.

In a third aspect, the present application provides an electronic device comprising:

a memory for storing program instructions;

A processor for invoking program instructions stored in the memory and executing the steps comprised by the method according to any of the first aspects in accordance with the obtained program instructions.

In a fourth aspect, the present application provides a computer readable storage medium storing a computer program comprising program instructions which, when executed by a computer, cause the computer to perform the method of any one of the first aspects.

In a fifth aspect, the present application provides a computer program product comprising: computer program code which, when run on a computer, causes the computer to perform the method of any of the first aspects.

The application has the following beneficial effects:

The embodiment of the application provides an asset configuration scheme generation method, an asset configuration scheme generation device, electronic equipment and a medium, wherein the method comprises the following steps: after the basic information and the asset information of the target object needing asset configuration are acquired, industry category preference prediction can be performed on the target object based on the basic information and the asset information, and the industry category preference information of the target object is determined. At least one target financial product may then be selected from the plurality of financial products included in the target financial product type based on the industry category preference information. Finally, outputting an asset allocation scheme indicating the proportion of each target financial product in the assets of the target object based on the selected at least one target financial product.

According to the method, firstly, the industry class preference of the target object can be predicted based on the basic information and the asset information of the target object, and the industry class preference of the target object can be predicted more accurately because the basic information of the user and the asset information of the user are considered. And then, based on the predicted industry class preference information, an asset configuration scheme which is more in line with the industry class preference of the target object can be provided for the target object, so that the problem of unstable quality of the asset configuration scheme generated in the related technology is avoided.

Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. The objectives and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.

Drawings

FIG. 1 is a schematic diagram of a system architecture to which an asset allocation scheme generation method according to an embodiment of the present application is applied;

FIG. 2 is one of exemplary flowcharts of a method for asset allocation scheme generation according to embodiments of the present application;

FIG. 3 is a flowchart of a training method of an industry category preference prediction model according to an embodiment of the present application;

FIG. 4 is a schematic diagram of a model structure according to an embodiment of the present application;

FIG. 5 is a flowchart illustrating a training method of a risk preference classification model according to an embodiment of the present application;

FIG. 6 is one of exemplary flowcharts of a method for asset allocation scheme generation according to embodiments of the present application;

FIG. 7 is a block diagram of an asset allocation scheme generation device according to an embodiment of the present application;

fig. 8 is a block diagram of an electronic device according to an embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application. Embodiments of the application and features of the embodiments may be combined with one another arbitrarily without conflict. Also, while a logical order of illustration is depicted in the flowchart, in some cases the steps shown or described may be performed in a different order than presented.

The terms first and second in the description and claims of the application and in the above-mentioned figures are used for distinguishing between different objects and not for describing a particular sequential order. Furthermore, the term "include" and any variations thereof is intended to cover non-exclusive protection. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus. The term "plurality" in the present application may mean at least two, for example, two, three or more, and embodiments of the present application are not limited.

In the technical scheme of the application, the data is collected, transmitted, used and the like, and all meet the requirements of national relevant laws and regulations.

Before describing the asset allocation scheme generating method provided by the embodiment of the present application, for convenience of understanding, the following detailed description will be given to the technical background of the embodiment of the present application.

With the continuous development of the financial market and the continuous improvement of the living standard of people, the purchasing demands of people for financial products are also increasing. Especially for high equity customers, such as customers using private banking services, it is important how to distribute their assets reasonably because of the large amount of own assets.

In order to avoid such a situation, the embodiment of the application provides an asset configuration scheme generating method, in which after basic information and asset information of a target object needing asset configuration are acquired, industry category preference prediction can be performed on the target object based on the basic information and the asset information, and industry category preference information of the target object is determined. At least one target financial product may then be selected from the plurality of financial products included in the target financial product type based on the industry category preference information. Finally, outputting an asset allocation scheme indicating the proportion of each target financial product in the assets of the target object based on the selected at least one target financial product.

Referring to fig. 1, a schematic diagram of a system architecture to which an asset allocation scheme generating method according to an embodiment of the present application is applied is shown. The system may be a server 100 comprising: memory 110, processor 120, and communication interface 130. The communication interface 130 may be configured to obtain basic information and asset information of the target object, and send the generated asset configuration scheme to the terminal device of the target object. The processor 120 is a control center of the server 100, connects various parts of the entire server 100 using various interfaces and lines, and performs various functions of the server 100 and processes data by running or executing software programs or modules stored in the memory 110, and calling data stored in the memory 110. Optionally, the processor 120 may include one or more processing units. The memory 110 may be a high-speed random access memory, or may be a non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device. It should be noted that the structure shown in fig. 1 is merely an example, and the embodiment of the present application is not limited thereto.

Referring to fig. 2, which is one exemplary flowchart of a method for generating an asset allocation scheme according to an embodiment of the present application, the method may be applied to the server shown in fig. 1, and may include the following procedures:

S201, basic information and asset information of a target object are acquired.

The target object may refer to an object that needs to be configured for an asset. The flow shown in fig. 2 may be triggered upon receipt of indication information sent by the target object indicating that an asset configuration operation is to be performed.

The basic information may include information characterizing basic features of the target object. Such as age, family condition, academic, profession, etc. The basic information can be input by the target object when relevant business is handled.

The asset information may include information characterizing asset conditions of the target object, as well as business processes. For example: asset management size (Assets Under Management, AUM), time of service registration, number of service transactions in a specified period, number of transactions in a specified period, and the like. Wherein the asset information may be historical data generated when the target object transacts the related business.

Alternatively, after the server obtains the basic information and the asset information, the derivative information of the target object may be obtained by performing operations such as calculation and comparison on the basic information and the asset information, and the derivative information may represent changes of some basic information or asset information, for example, an AUM annual growth rate, a family condition change, and the like.

Before the basic information and the asset information are acquired, the target object may be queried by sending a message to the target object, or the like, whether the server is allowed to acquire and use the basic information and the asset information, and the flow shown in fig. 2 may be executed after the target object agrees.

In some embodiments, the target object may also be an object that satisfies a preset condition. For example, to assist some objects in better optimizing their asset configuration structure, the target object may also be a low liveness object whose asset configuration liveness is not above a preset liveness threshold. Asset allocation liveness may represent, to some extent, how often an object is asset allocated.

Alternatively, the asset allocation liveness may be determined according to a ratio between an asset management scale of the object's specified financial product type asset and a total asset scale of the object, satisfying equation (1).

S=1-AUM _c/AUM_a formula (1)

Where S represents the asset allocation liveness, AUM _c represents the asset management scale of the specified category of asset of the object, and AUM _a represents the total asset scale of the object. The type of the asset specifying the type of financial product may be determined based on actual conditions or experience, and may be, for example, a fixed revenue-type asset or a deposit-type asset, as the application is not limited in this regard.

S202, based on the basic information and the asset information, industry category preference prediction is conducted on the target object, and the industry category preference information of the target object is determined.

The industry category preference information characterizes preference degree of the target object on the industry category to which the financial product belongs. Industry categories may include finance, real estate, industry, information technology, alternative consumption, healthcare, utilities, energy, and the like.

In a possible implementation manner, the server predicts the industry category preference of the target object based on the basic information and the asset information of the target object, and when determining the industry category preference information of the target object, the server may perform feature extraction on the basic information and the asset information of the target object to determine a feature set of the target object. And analyzing the feature set of the target object based on the association relation between the feature set of the different objects and the industry category preference of the different objects to obtain the industry category preference information of the target object.

In a possible implementation manner, the server performs feature extraction on basic information and asset information of the target object, and when determining a feature set of the target object, the server may perform feature extraction on the basic information and the asset information first to obtain a candidate feature set of the target object. One or more features are then screened from the candidate feature set to form a feature set. Wherein, the characteristics in the characteristic set satisfy: the one or more features are associated with industry class preferences to a greater extent than other features in the candidate feature set are associated with industry class preferences.

In some embodiments, when the server screens one or more features from the candidate feature set to form the feature set, random features can be generated based on each candidate feature included in the candidate feature set, and candidate features with influence degree on the model variable being greater than that of the random features can be used as undetermined features. And constructing a classification model aiming at industry class preference through an automatic modeling model, and sequencing importance degrees of all undetermined features based on the influence degree of the undetermined features on the model output accuracy. And finally, determining a plurality of features included in the feature set based on the sorting result. Where model variables refer to variables used to train models of random and candidate features.

In one example, in determining the pending features in the candidate feature set, the candidate features may be initially screened by a Boruta algorithm. The Boruta algorithm is a wrapper method based on a random forest classification algorithm, and the Boruta algorithm randomly arranges candidate features to generate a set of random features, and then inputs the random features and the candidate features into a random forest model for training. The relative importance of each candidate feature may then be determined by comparing the importance of the candidate feature with the importance of the random feature. Features of high importance are taken as pending features. Wherein, the influence degree of the feature with high importance on the random forest model variable is larger. For example, a candidate feature is considered to be of high importance if it is significantly more important than the corresponding random feature.

In one example, each of the pending features may be input into an Automated modeling model, such as an Automated machine learning (Automated MACHINE LEARNING, AUTOML) model, when the features belonging to the feature set are determined. Automatic modeling can be performed by AutoML models with industry class preference classification as a target, and a plurality of models can be trained in limited resources simultaneously by AutoGluon toolkits in the models and fused together to realize automatic machine learning. After modeling, determining the importance degree of the undetermined features based on the influence degree of each undetermined feature on the output accuracy of the constructed model. And sequencing the undetermined features from high importance level to low importance level. When determining the plurality of features included in the feature set based on the ranking result, it may be determined that the undetermined feature whose ranking is in the first 80% is the plurality of features included in the feature set. It should be noted that the feature selection ratio may be set according to actual situations or experience, which is not limited by the present application.

Based on the above scheme, assuming 241 candidate features can be obtained based on the basic information and the asset information, 197 pending features can be obtained after the primary screening, and 157 features are included in the possible final feature set after the secondary screening. Therefore, less important characteristics can be removed through multiple screening, and interference of the characteristics on output results is avoided, so that predicted industry category preference is more accurate. And by reducing the number of features, the processing speed in analyzing the feature set of the target object to obtain industry category preference information of the target object can be increased.

In some embodiments, when the feature set of the target object is analyzed to obtain the industry category preference information of the target object based on the association relationship between the feature set of the different object and the industry category preference of the different object, the industry category preference prediction of the target object can be performed according to the feature set by using the trained industry category preference prediction model, so as to determine the industry category preference information of the target object. The industry class preference prediction model is trained based on a feature set of a training sample and an industry class preference label of the training sample, wherein the training sample is an object with asset configuration liveness greater than a preset liveness threshold.

Referring to fig. 3, a flowchart of a training method of an industry category preference prediction model according to an embodiment of the present application is shown, where the flowchart includes:

s301, acquiring a training sample set.

Specifically, before training the industry class preference prediction model, the server may determine, through formula (1), the asset configuration liveness of each of the plurality of objects, and use, as a training sample, an object whose asset configuration liveness is greater than a preset liveness threshold.

The preset activity threshold may be a fixed value, for example, 0.7 or 0.8, which is preset. Or the preset liveness threshold value may be determined according to the sorting result after sorting the plurality of objects according to the asset configuration liveness. For example, when 500 objects are ranked according to the asset allocation liveness and the first 10% of objects are required to be used as training samples, the asset allocation liveness of the objects ranked at the 50 th position is used as a preset liveness threshold.

In a possible implementation manner, the server may determine the asset configuration liveness of each of the plurality of objects through the formula (1), and take the object with the asset configuration liveness being greater than the preset liveness threshold as the candidate training sample. And then, acquiring basic information and asset information of a plurality of candidate training samples, carrying out statistical analysis on the basic information and the asset information, and taking the candidate training samples with the information deletion ratio not larger than a preset deletion ratio threshold as training samples. The preset deletion proportion threshold may be set empirically or in actual situations, for example, may be 20%, 10%, etc., which is not limited in the present application. The method for determining the candidate training samples based on the configuration liveness of the asset may be referred to as the method for determining the training samples based on the configuration liveness, which is not described herein.

For example, assume that basic information and asset information include the following seven items of information: age, family condition, academic, professional, AUM, family change condition, business transaction times in one year. The information deletion ratio of the object A is about 29% when two pieces of information, namely occupation and family change conditions, are deleted from the basic information and the asset information of the object A; the three information of occupation, family change condition and business handling times in one year are deleted in the basic information and the asset information of the object B, and the information deletion proportion of the object B is about 43%; the basic information and the asset information of the object C are not missing any information, and the information missing proportion of the object C is 0%; the information deletion rate of the object D is about 14% when the information of the family change is deleted from the basic information and the asset information of the object D. Assuming that the preset deletion percentage threshold is 30%, since only the information deletion percentage of the object B is greater than 30% in the objects a to D, the objects a, C, and D can be regarded as training samples.

Because part of the object service registration time is long, some information may be lost, so that samples can be screened before training a model. After the samples are screened by the scheme, the situation that the trained model output result is inaccurate due to excessive information loss can be avoided.

In one possible implementation, after the training samples are determined, the information about the missing information in the training samples may be filled in by means of median or mean. For example, when the asset information of the training sample E is filled with the median, the information of "the number of business processes in one year" is deleted, and if the median of the information of "the number of business processes in one year" in each training sample is 10, "10" is taken as the "number of business processes in one year" of the training sample E.

After the training samples are determined and filled with information, feature extraction can be performed on basic information and asset information of each training sample respectively, and a feature set of each training sample is determined. The feature extraction method may be referred to the related description in the above method embodiments, which is not described herein.

In one possible implementation, the features included in the feature set may be determined before the industry class preference prediction model is trained, and then when feature extraction is performed on the basic information and the asset information of the target object, feature extraction may be performed based on the features included in the determined feature set, without performing the feature screening operation again.

For example, determining the feature set prior to training the industry class preference prediction model includes: age, family condition, academic, professional, AUM, family change condition, business transaction times in one year. When the feature extraction is performed on the target object, the feature extraction can be performed directly on the information of the target object, so as to obtain a feature set of the target object.

In some embodiments, for each training sample, the held financial product of the training sample may be determined based on the asset information of the training sample, thereby determining the industry category to which the financial product belongs. The industry class preference tags for the training samples may then be determined based on ordering the holding amounts for financial products belonging to different industry classes.

For example, assume that the financial products held by the training sample include a financial product A1, a financial product A2, a financial product B1, and a financial product C1, wherein the industry category to which the financial product A1 and the financial product A2 belong is finance, the industry category to which the financial product B1 belongs is obligatory consumption, and the industry category to which the financial product C1 belongs is energy. The industry category to which the financial product held by the training sample belongs can be determined to comprise finance, obligatory consumption and energy, the total holding amount of the financial product A1 and the financial product A2 is used as the holding amount corresponding to finance, the holding amount of the financial product B1 is used as the holding amount corresponding to obligatory consumption, and the holding amount of the financial product C1 is used as the holding amount corresponding to energy. The financial, alternative consumption and energy corresponding holding amounts may then be ranked such that one or more industry categories thereof are used as industry category preference tags for the training sample.

The number of industry category preferences included in the industry category preference tag may be set according to actual situations or experiences, which is not limited in the present application. For example, the industry category with the top 3 monetary amount per training sample may be determined as the industry category preference label for that training sample. Or the industry category with the highest sum held by each training sample can be used as the industry category preference label of the training sample.

S302, training sample balance processing.

In order to make the output result of the trained industry class preference prediction model more accurate, the number of training samples corresponding to each class of industry class preference labels needs to be balanced as much as possible.

Specifically, the server may perform sample expansion processing on industry class preference tags with a smaller number of samples using a synthetic minority class oversampling technique (SYNTHETIC MINORITY OVER-sampling Technique, SMOTE) algorithm. The specific implementation can be carried out by the following steps:

1. and determining the industry class preference label to be expanded with a small sample number.

Industry category preference tags with a small number of samples may be determined according to preset rules. For example, a sample number threshold may be set, and an industry category preference tag with a sample number less than the sample number threshold is the industry category preference tag to be expanded.

For another example, a sample number difference between every two industry category preference tags may be calculated, and when the sample number difference is greater than a preset difference threshold, the sample number of the two industry category preference tags may be considered to be the industry category preference tag to be expanded. For example, assume that there are 500 training samples with industry class preference labels A, 200 training samples with industry class preference labels B, 100 training samples with industry class preference labels C, and a difference threshold of 200. And if the difference value of the sample number between A and B is larger than the difference value threshold value and the sample number of B is smaller, the B can be considered as the industry class preference label to be expanded. And if the difference value of the sample number between A and C is larger than the difference value threshold value and the sample number of C is smaller, then C can be regarded as the industry class preference label to be expanded.

2. And determining neighbor samples of the training samples belonging to the industry class preference labels to be expanded.

For each training sample belonging to the industry class preference label to be expanded, the following operations can be performed separately: and calculating Euclidean distances between the training samples and other training samples belonging to the same industry class preference label to be expanded, thereby determining k neighbor samples of the training samples. k is a positive integer, which can be set according to practical situations, and the application is not limited thereto.

3. And determining the sampling multiplying power.

The sampling ratio may be set according to the sample imbalance ratio to determine the sampling magnification N. Wherein the sampling rate is the ratio between the number of new training samples generated and the number of original minority class samples. The sampling rate may generally be an integer greater than 1.

For example, if the number of samples belonging to the industry class preference label to be expanded is 100 and the sampling magnification is set to 2, the number of new training samples generated will be twice the number of original samples, i.e., 200.

4. A random number between 0 and 1 is generated.

5. A new training sample is constructed.

For each training sample belonging to the industry class preference label to be expanded, randomly selecting a plurality of samples from k adjacent samples, and for each randomly selected adjacent sample, respectively constructing a new training sample with the original training sample according to a formula (2).

X _new＝x+rand(0,1)*(X_j -X), j=1, 2,3, …, n formula (2)

Wherein, X _new represents a new training sample, X belongs to a training sample of the industry class preference label to be expanded, rand (0, 1) represents a random number between 0 and 1, X _j represents a neighbor sample of X, and j can be taken from 1 to n.

S303, training an industry category preference prediction model based on the feature set of the training sample and the industry category preference label.

The industry class preference prediction model may use Long Short-Term Memory (LSTM), and fig. 4 is a schematic diagram of a model structure provided by an embodiment of the present application. The LSTM model may include a random inactivation layer (Drop-out), a full connection layer (Dense), and a softMax layer. Inputting the feature set of the training sample and the industry class preference label into the LSTM model, avoiding overfitting through a Drop-out layer, outputting the prediction result of the industry class preference of the training sample through a softMax layer, and then adjusting the parameters of the LSTM model based on the loss between the prediction result and the industry class preference label until the trained industry class preference prediction model is obtained when the preset training ending condition is met.

The preset training ending condition may be that the loss function converges, or that the maximum iteration number is reached, etc., and the training ending condition is not limited by the present application. The loss function can be set according to actual conditions or experience, and the selection of the loss function is not limited in the application.

Optionally, the training samples may be further divided into a training set, a test set, and a validation set according to an 8:1:1 ratio. The industry class preference prediction model is trained through the feature sets and the industry class preference labels of all training samples in the training set, the industry class preference prediction model is tested according to the feature sets and the industry class preference labels of all training samples in the test set after training, and finally the industry class preference prediction model can be evaluated based on the feature sets and the industry class preference labels of all training samples in the verification set. And obtaining a trained industry class preference prediction model when the model index reaches a preset condition. Model metrics may include F1 score, area under ROC curve (Area Under the Curve, AUC), and the like. It should be appreciated that the method of testing and verifying the model may take the form of methods in the related art, as the application is not limited in this regard.

S203, selecting at least one target financial product from a plurality of financial products included in the target financial product type according to the industry category preference information.

In a possible implementation, before executing S203, the target risk preference of the target object may also be determined according to the basic information and the asset information of the target object.

Specifically, the server may perform feature extraction on the basic information and the asset information, and determine a feature set of the target object. The feature set of the target object is analyzed to obtain the risk preferences of the target object based on the association between the feature set of the different object and the risk preferences of the different object.

In some embodiments, when the server analyzes the feature set of the target object to obtain the risk preference of the target object based on the association relationship between the feature set of the different object and the risk preference of the different object, the server may classify the risk preference of the target object according to the feature set by using a trained risk preference classification model, so as to determine the target risk preference of the target object. The risk preference classification model is trained based on a feature set of training samples and risk preference labels, wherein the training samples are objects with liveness greater than a preset liveness threshold; liveness characterizes the liveness of an object for asset configuration.

Fig. 5 is a schematic flow chart of a training method of a risk preference classification model according to an embodiment of the present application, where the flow includes:

s501, acquiring a training sample set.

The method for determining the training samples is the same as the method for determining the training samples before training the industry class preference prediction model, and reference may be made to the description of the correlation in S301, which is not repeated here.

In one possible implementation manner, when determining the feature set of the training sample, the server may determine a candidate feature set according to the feature screening method described in S202, and then screen one or more features from the candidate feature set to form the feature set. Wherein, the characteristics in the characteristic set satisfy: the one or more features are associated with the risk preference to a greater extent than the other features in the candidate feature set.

The specific method of feature screening can participate in the related description in the embodiment of the method, and when the method is different from the method of feature screening, the method can automatically model by taking risk preference classification as a target when the classification model is constructed by the automatic modeling model, so that a plurality of features included in the feature set are determined, and are not repeated herein.

In another possible implementation manner, when the one or more features are selected from the candidate feature set to form the feature set, the features in the feature set may further satisfy that the association degree of the one or more features with the risk preference is greater than the association degree of the other features in the candidate feature set with the risk preference, and the association degree of the one or more features with the industry class preference is greater than the association degree of the other features in the candidate feature set with the industry class preference.

In specific implementation, the related characteristics of the industry category preference can be screened out according to the related degree of the characteristics and the industry category preference by the characteristic screening method described in the step S202. And constructing a classification model by taking risk preference classification as a target through an automatic modeling model, inputting the screened industry class preference associated features into the classification model, thereby obtaining importance degree ordering of the features, and determining each feature included in the feature set.

It should be noted that, when determining the feature set of the target object, the above method may also be used to perform screening, that is, one or more features are screened from the candidate feature set, and when forming the feature set, the features in the feature set may satisfy: the one or more features are more correlated with the risk preferences than other features in the candidate feature set, or the one or more features are more correlated with the industry category preferences than other features in the candidate feature set, or the one or more features are more correlated with the risk preferences than other features in the candidate feature set, and the one or more features are more correlated with the industry category preferences than other features in the candidate feature set.

In some embodiments, for each training sample, the held financial product of the training sample may be determined based on the asset information of the training sample, thereby determining the risk level to which the financial product belongs. The risk preference score for the training sample may then be determined by calculation based on the holding amounts of financial products belonging to different risk levels through equation (3).

Where Score represents risk preference Score of the training sample, x represents risk level, p represents the number of preset risk levels, SUM _x represents the amount of financial product held by the training sample belonging to the x risk level, and SUM _a represents the total amount of financial product held by the training sample.

For example, assuming that the risk levels to which the financial products belong are divided into 1, 2,3, 4, and 5, the risk preference score of the training sample may be determined according to the amount of the financial products to which the risk level held by the training sample belongs to 1, the amount of the financial products to which the risk level held by the training sample belongs to 2, the amount of the financial products to which the risk level held by the training sample belongs to 3, the amount of the financial products to which the risk level held by the training sample belongs to 4, and the amount of the financial products to which the risk level held by the training sample belongs to 5.

After determining the risk preference score for each training sample, the risk preference score for each training sample may be normalized according to the mean and standard deviation of the risk preference scores, satisfying equation (4).

Where Score' represents the standard value of the risk preference Score for the training sample, μ represents the average value of the risk preference scores for each training sample, σ represents the standard deviation of the risk preference scores for each training sample.

After normalizing the risk preference score, the risk preference label of the training sample with the standard value of the risk preference score equal to or greater than sigma may be marked as 1, the risk preference label of the training sample with the standard value of the risk preference score < sigma is marked as 0, and the risk preference label of the training sample with the standard value of the risk preference score equal to or less than-sigma is marked as-1.

S502, training sample balance processing.

In order to make the output result of the trained risk preference classification model more accurate, the number of training samples corresponding to each type of risk preference label needs to be balanced as much as possible. The specific method is similar to the sample balance processing method in training the industry class preference prediction model, and can participate in the related description in S302, which is not described in detail herein.

S503, training a risk preference classification model based on the feature set of the training sample and the risk preference label.

The risk preference classification model may also employ an LSTM model, the model structure of which is shown in fig. 4. Parameters of the LSTM model can be adjusted based on losses between the prediction classification result output by the LSTM and the risk preference labels until a trained risk preference classification model is obtained when preset training ending conditions are met.

The preset training ending condition may be the same as or different from the preset training ending condition of the industry class preference prediction model, which is not limited in the present application. The loss function may be the same as or different from the preset training ending condition of the industry class preference prediction model, which is not limited in the present application.

Optionally, the training samples may be further divided into a training set, a test set, and a validation set according to an 8:1:1 ratio. And training a risk preference classification model through the feature set and the risk preference label of each training sample in the training set, testing the risk preference classification model according to the feature set and the risk preference label of each training sample in the test set after training, and finally evaluating the risk preference classification model based on the feature set and the risk preference label of each training sample in the verification set. And obtaining a trained risk preference classification model when the model index reaches a preset condition. Model metrics may include F1 metrics, AUC metrics, and the like. It should be appreciated that the method of testing and verifying the model may take the form of methods in the related art, as the application is not limited in this regard.

In some embodiments, after determining the target risk preference of the target object, the server may also select a target financial product type for the target object based on the target risk preference of the target object prior to performing S203. Wherein the target risk preference characterizes a risk acceptance of the target object. The financial product types may include fixed returns, equity, money, merchandise, security, etc., the fixed returns may include financial products such as regular deposits, the equity may include financial products such as stocks, funds, etc., the merchandise may include financial products of a physical type such as precious metals, etc., and the security may include financial products such as insurance.

The server may select a target financial product belonging to the target industry category among the plurality of financial products included in the target financial product type according to the target industry category recorded in the industry category preference information when selecting at least one target financial product from the plurality of financial products included in the target financial product type according to the industry category preference information.

For example, the target financial product type of the target object may include A, B, and the server selects the financial product belonging to the target industry category as the target financial product of the target object from the plurality of financial products A1, A2, A3 … … belonging to the a according to the industry category preference information, and then selects the financial product belonging to the target industry category as the target financial product of the target object from the plurality of financial products B1, B2, B3 … … belonging to the B.

The number of the selected target financial products belonging to the target industry category may be set according to actual situations, for example, when there are a plurality of target industry categories, one target financial product may be selected for each target industry category. Multiple target financial products may also be selected separately for each target industry category. The application is not limited in this regard.

In some embodiments, in the plurality of financial products included in the target financial product type, there may be a plurality of financial products belonging to the target industry category, and then the plurality of financial products belonging to the target industry category may be ranked based on the preset evaluation value of each financial product, and the financial product ranked in front may be selected as the target financial product.

For example, each of the financial products A-D belongs to the target industry category, the evaluation value of the financial product A is 80, the evaluation value of the financial product B is 90, the evaluation value of the financial product C is 85, and the evaluation value of the financial product D is 88. If the number of target financial products is 3, the financial product B, C, D may be determined as the target financial product. If the number of target financial products is 2, the financial product B, D may be determined as the target financial product.

In a possible implementation manner, before the asset configuration scheme generating method provided by the embodiment of the application is executed, the risk preference levels of the objects with the liveness greater than the preset liveness threshold value can be determined through a formula (3) and a formula (4). At each risk preference level, statistics are made on the asset management scale of the object, respectively, and a plurality of asset management scale levels are divided. And counting the proportion of the total asset occupied by the amount of each financial product type held by the object according to any risk preference level and any asset management scale level under the risk preference level to obtain the mapping relation between the asset management scale level under each risk preference level and the proportion of the financial product type.

For example, the asset management scale levels may include 0-10000, 10000-100000, 100000-1000000, 1000000-10000000, 10000000 or more, five levels.

When the risk preference level includes 1,0, -1, the server may determine a mapping relationship between each asset management scale level and a financial product type ratio when the risk preference level is1, a mapping relationship between each asset management scale level and a financial product type ratio when the risk preference level is 0, and a mapping relationship between each asset management scale level and a financial product type ratio when the risk preference level is-1, respectively.

Taking the risk preference level of 1 and the asset management scale level of 0-10000 as an example for explanation, the objects with the risk preference level of 1 and the asset management scale level of 0-10000 can be counted, the proportion of the sum of the held financial product types to the total asset is determined, and the average value of the proportion of the objects is used. And when the risk preference level is determined to be 1, the mapping relation between the asset management scale level 0-10000 and the financial product type ratio is determined.

For example, assuming that the financial product type includes A, B, C, the object with a risk preference level of 1 and an asset management scale level of 0-10000 includes: object 1-object 3. Wherein, the proportion of A in the asset of the object 1 is 20%, the proportion of B is 30%, the proportion of C is 50%, the proportion of A in the asset of the object 2 is 40%, the proportion of B is 30%, the proportion of C is 30%, the proportion of A in the asset of the object 3 is 20%, the proportion of B is 40%, and the proportion of C is 40%. Then, according to the average value of the duty ratios of the same financial product types of the objects 1-3, it can be determined that in the case of the risk preference level of 1, the asset management scale level 0-10000 corresponds to the financial product type duty ratio, where a is about 27%, B is about 33%, and C is about 40%. Other risk preference levels and other mapping relationships between asset management scale levels and financial product type ratios can be calculated by the same method, and will not be described in detail herein.

After determining the mapping relation between each asset management scale level and the ratio of the financial product types under each risk preference level, when selecting the target financial product type for the target object based on the risk preference of the target object, the server can determine the mapping relation between the asset management scale level applicable to the target object and the ratio of the financial product types, and then determine the asset management scale level to which the target object belongs based on the asset management scale of the target object, thereby determining the financial product type corresponding to the target object and the corresponding ratio of the financial product types.

S204, outputting an asset allocation scheme based on the at least one target financial product.

Wherein the asset allocation scheme indicates a specific gravity of each of the target financial products in the assets of the target object.

In one implementation, after determining the target financial product corresponding to the target object, the server may determine a specific gravity of the target financial product in the asset of the target object based on a specific gravity of the financial product type corresponding to the target object.

For example, assume that the target object corresponds to a financial product type with a 40% duty cycle of a, a 30% duty cycle of B, and a 30% duty cycle of C. The target financial products of the financial product type A comprise A1, A2 and A3, the target financial products of the financial product type B comprise B1, B2 and B3, and the target financial products of the financial product type C comprise C1, C2 and C3. Then A1, A2, A3 would be 40% of the sum of the specific gravities of the target object's assets, B1, B2, B3 would be 30% of the sum of the specific gravities of the target object's assets, and C1, C2, C3 would be 30% of the sum of the specific gravities of the target object's assets.

Alternatively, the specific gravity of each target financial product in the asset of the target object may be determined according to a preset evaluation value of each target financial product. Or the specific gravity of each target financial product in the assets of the target object may be evenly distributed according to the specific gravity of the financial product type to which the target financial product belongs.

In one example, assuming that the target financial product includes A1, A2, A3, B1, B2, B3, C1, C2, C3, the asset configuration scheme of the output target object may include: a1 accounts for 14 percent, A2 accounts for 13 percent, A3 accounts for 13 percent, B1 accounts for 10 percent, B2 accounts for 10 percent, B3 accounts for 10 percent, C1 accounts for 10 percent, C2 accounts for 10 percent and C3 accounts for 10 percent.

In the following, in order to more clearly understand the solution provided by the embodiments of the present application, an asset configuration solution generating method provided by the present application will be described with reference to specific embodiments.

Referring to fig. 6, one of exemplary flowcharts of an asset allocation scheme generating method according to an embodiment of the present application specifically includes:

s601, information preprocessing.

Basic information and asset information of each of a plurality of objects are acquired, training samples are determined, and features included in a feature set of each training sample are determined. The method for determining the training samples may be referred to the relevant description in the method embodiment shown in fig. 3 or fig. 5, and the method for determining the features may be referred to the relevant description in the method embodiment shown in fig. 2, fig. 3 or fig. 5, which are not described herein.

S602, training a risk preference classification model and an industry category preference prediction model.

The server may train the risk preference classification model through the process shown in fig. 5 and the industry category preference prediction model through the process shown in fig. 3.

S603, generating an asset configuration scheme.

Basic information and asset information of the target object are acquired, and feature extraction is performed on the basic information and the asset information of the target object based on the features included in the feature set determined in the step S601, so as to obtain a feature set of the target object. And respectively inputting the feature set of the target object into a trained risk preference classification model and an industry category preference prediction model to obtain target risk preference and industry category preference information of the target object. And then generating an asset configuration scheme of the target object based on the target risk preference and industry category preference information of the target object, and transmitting the generated asset configuration scheme to the target object.

The asset configuration scenario generation method may be referred to in the related description of the method embodiment shown in fig. 2, and will not be described herein.

Based on the same inventive concept, an embodiment of the present application provides an asset allocation scheme generating apparatus, please refer to fig. 7, which includes:

A communication unit 701, configured to acquire basic information and asset information of a target object;

a preference information determining unit 702, configured to predict an industry category preference of the target object based on the basic information and the asset information, and determine industry category preference information of the target object, where the industry category preference information characterizes a preference degree of the target object for an industry category to which the financial product belongs;

An asset allocation unit 703 for selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information; an asset allocation scheme is output based on the at least one target financial product, the asset allocation scheme indicating a specific gravity of each target financial product in the assets of the target object.

In a possible implementation manner, before the asset configuration unit 703 selects at least one target financial product from the plurality of financial products included in the target financial product type according to the industry category preference information, the asset configuration unit is further configured to: selecting a target financial product type for the target object based on the risk preference of the target object; the target risk preference characterizes a risk acceptance of the target object; the asset allocation unit is configured to, when selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information: and selecting a target financial product belonging to the target industry category from a plurality of financial products included in the target financial product type according to the target industry category recorded in the industry category preference information.

In a possible implementation manner, the target risk preference of the target object is determined by the preference information determining unit 702 according to the following method: extracting features of the basic information and the asset information, and determining a feature set of the target object; analyzing the feature set of the target object based on the association relation between the feature set of the different objects and the risk preferences of the different objects to obtain the risk preferences of the target object.

In a possible implementation manner, the preference information determining unit 702 is configured to, when analyzing the feature set of the target object to obtain the risk preference of the target object based on an association relationship between the feature set of the different object and the risk preference of the different object: performing risk preference classification on the target object according to the feature set through a trained risk preference classification model, and determining target risk preference of the target object; the risk preference classification model is trained based on feature sets and risk preference labels of training samples, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

In a possible implementation manner, the preference information determining unit 702 performs industry category preference prediction on the target object based on the basic information and the asset information, and when determining industry category preference information of the target object, is used to: extracting features of the basic information and the asset information, and determining a feature set of the target object; and analyzing the characteristic set of the target object based on the association relation between the characteristic set of the different objects and the industry category preference of the different objects to obtain the industry category preference information of the target object.

In a possible implementation manner, the preference information determining unit 702 is configured to, when analyzing the feature set of the target object to obtain the industry category preference information of the target object based on an association relationship between the feature set of the different object and the industry category preference of the different object: carrying out industry category preference prediction on the target object according to the feature set by using a trained industry category preference prediction model, and determining industry category preference information of the target object; the industry class preference prediction model is trained based on a feature set of training samples and an industry class preference label, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

In a possible implementation manner, the preference information determining unit 702 performs feature extraction on the basic information and the asset information, and when determining a feature set of the target object, is configured to: extracting features of the basic information and the asset information to obtain a candidate feature set of the target object; screening one or more features from the candidate feature set to form the feature set; wherein, the characteristics in the characteristic set satisfy: the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set or the one or more features are associated with an industry category preference that is greater than the other features in the candidate feature set or the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set and the industry category preference that is greater than the other features in the candidate feature set.

In a possible implementation manner, the asset configuration unit 703 is configured to screen one or more features from the candidate feature set, and when forming the feature set, is configured to: based on each candidate feature included in the candidate feature set, respectively generating random features, and taking the candidate features with influence degree on model variables being greater than that of the random features as undetermined features; constructing a classification model through an automatic modeling model, and sequencing importance degrees of all undetermined features based on the influence degree of the undetermined features on the model output accuracy; the classification model is a model for classifying industry category preferences and/or risk preferences; based on the ranking result, a plurality of features included in the feature set are determined.

Based on the same concept of the above method, referring to fig. 8, a schematic structural diagram of an electronic device according to an embodiment of the present application is provided, where the electronic device includes at least one processor 802, and a memory 801 connected or coupled to the at least one processor 802, and further, the electronic device may further include a communication interface 803. The electronic device may interact with other devices via communication interface 803.

By way of example, the communication interface 803 may be a transceiver, a circuit, a bus, a module, a pin, or other type of communication interface. When the electronic device is a chip-type device or circuit, the communication interface 803 in the electronic device may also be an input/output circuit, and may input information (or called receiving information) and output information (or called transmitting information), and the processor may be an integrated processor or a microprocessor or an integrated circuit or a logic circuit, and the processor may determine the output information according to the input information.

The coupling in the embodiments of the present application is an indirect coupling or communication connection between devices, units, or modules, which may be in electrical, mechanical, or other forms for information interaction between the devices, units, or modules. The processor 802 may cooperate with the memory 801 and the communication interface 803. The specific connection medium between the processor 802, the memory 801, and the communication interface 803 is not limited in the present application.

Optionally, referring to fig. 8, the processor 802, the memory 801 and the communication interface 803 are connected to each other through a bus. The bus may be a peripheral component interconnect standard (PERIPHERAL COMPONENT INTERCONNECT, PCI) bus, or an extended industry standard architecture (extended industry standard architecture, EISA) bus, or the like. The buses may be classified as address buses, data buses, control buses, etc. For ease of illustration, only one arrow is shown in FIG. 8, but not only one bus or type of bus.

In an embodiment of the present application, the memory 801 is used as a non-volatile computer readable storage medium for storing non-volatile software programs, non-volatile computer executable programs, and modules. The Memory 801 may include at least one type of storage medium, and may include, for example, flash Memory, hard disk, multimedia card, card Memory, random access Memory (Random Access Memory, RAM), static random access Memory (Static Random Access Memory, SRAM), programmable Read-Only Memory (Programmable Read Only Memory, PROM), read-Only Memory (ROM), charged erasable programmable Read-Only Memory (ELECTRICALLY ERASABLE PROGRAMMABLE READ-Only Memory, EEPROM), magnetic Memory, magnetic disk, optical disk, and the like. Memory 801 is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto. The memory 801 in the embodiments of the present application may also be circuitry or any other device capable of implementing a storage function for storing instructions, computer programs, and/or data.

In an embodiment of the present application, the processor 802 may be a general-purpose processor, such as a Central Processing Unit (CPU), digital signal processor, application specific integrated circuit, field programmable gate array or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, and may implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present application. The general purpose processor may be a microprocessor or any conventional processor or the like. The steps of the asset allocation scheme generation method disclosed in connection with the embodiment of the application can be directly embodied as a hardware processor or a combination of hardware and software modules in the processor.

By programming the processor 802, the code corresponding to the asset allocation scheme generating method described in the foregoing embodiment may be cured into the chip, so that the chip can execute the steps of the foregoing asset allocation scheme generating method during operation, and how to program the processor 802 is a technology known to those skilled in the art, which is not repeated herein.

In one or more embodiments, the memory 801 stores instructions executable by the at least one processor 802, and the at least one processor 802 may implement steps of any of the methods described above by invoking instructions or computer programs stored in the memory 801.

Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, the computer program product comprising: computer program code which, when run on a computer, causes the computer to perform the asset allocation scheme generation method as any of the preceding discussion. Since the principle of the solution of the problem of the computer readable storage medium is similar to that of the asset allocation scheme generation method, the implementation of the computer readable storage medium may refer to the implementation of the method, and the repetition is omitted.

Based on the same inventive concept, embodiments of the present application also provide a computer program product comprising: computer program code which, when run on a computer, causes the computer to perform the asset allocation scheme generation method as any of the preceding discussion. Since the principle of the solution of the problem of the computer program product is similar to that of the asset allocation scheme generation method, the implementation of the computer program product may refer to the implementation of the method, and the repetition is omitted.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of user operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit or scope of the application. Thus, it is intended that the present application also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

1. A method of asset allocation scheme generation, the method comprising:

basic information and asset information of a target object are acquired;

2. The method of claim 1, wherein prior to selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information, the method further comprises:

Selecting a target financial product type for the target object based on the risk preference of the target object; the target risk preference characterizes a risk acceptance of the target object;

The selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information includes:

And selecting a target financial product belonging to the target industry category from a plurality of financial products included in the target financial product type according to the target industry category recorded in the industry category preference information.

3. The method of claim 2, wherein the target risk preference of the target object is determined according to the following method:

Extracting features of the basic information and the asset information, and determining a feature set of the target object;

Analyzing the feature set of the target object based on the association relation between the feature set of the different objects and the risk preferences of the different objects to obtain the risk preferences of the target object.

4. A method according to claim 3, wherein the analyzing the feature set of the target object to obtain the risk preference of the target object based on the association between the feature set of the different object and the risk preference of the different object comprises:

Performing risk preference classification on the target object according to the feature set through a trained risk preference classification model, and determining target risk preference of the target object;

the risk preference classification model is trained based on feature sets and risk preference labels of training samples, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

5. The method of claim 1, wherein the performing industry category preference prediction on the target object based on the base information and the asset information, determining industry category preference information for the target object, comprises:

And analyzing the characteristic set of the target object based on the association relation between the characteristic set of the different objects and the industry category preference of the different objects to obtain the industry category preference information of the target object.

6. The method of claim 5, wherein analyzing the feature set of the target object to obtain the industry category preference information of the target object based on the association between the feature set of the different object and the industry category preference of the different object, comprises:

Carrying out industry category preference prediction on the target object according to the feature set by using a trained industry category preference prediction model, and determining industry category preference information of the target object;

the industry class preference prediction model is trained based on a feature set of training samples and an industry class preference label, wherein the training samples are objects with asset configuration liveness greater than a preset liveness threshold.

7. The method according to any one of claims 3-6, wherein the feature extracting the basic information and the asset information to determine a feature set of the target object includes:

Extracting features of the basic information and the asset information to obtain a candidate feature set of the target object;

screening one or more features from the candidate feature set to form the feature set;

Wherein, the characteristics in the characteristic set satisfy: the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set or the one or more features are associated with an industry category preference that is greater than the other features in the candidate feature set or the one or more features are associated with a risk preference that is greater than the other features in the candidate feature set and the industry category preference that is greater than the other features in the candidate feature set.

8. The method of claim 7, wherein the screening one or more features from the candidate feature set to form the feature set comprises:

Based on each candidate feature included in the candidate feature set, respectively generating random features, and taking the candidate features with influence degree on model variables being greater than that of the random features as undetermined features;

Constructing a classification model through an automatic modeling model, and sequencing importance degrees of all undetermined features based on the influence degree of the undetermined features on the model output accuracy; the classification model is a model for classifying industry category preferences and/or risk preferences;

Based on the ranking result, a plurality of features included in the feature set are determined.

9. An asset allocation scheme generation device, the device comprising:

10. The apparatus of claim 9, wherein the asset allocation unit is further configured to, prior to selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information:

The asset allocation unit is configured to, when selecting at least one target financial product from a plurality of financial products included in a target financial product type according to the industry category preference information:

11. The apparatus according to claim 10, wherein the target risk preference of the target object is determined by the preference information determining unit according to the following method:

12. The apparatus according to claim 11, wherein the preference information determining unit is configured to, when analyzing the feature set of the target object to obtain the risk preference of the target object, based on an association relationship between the feature set of the different object and the risk preference of the different object:

13. The apparatus according to claim 9, wherein the preference information determining unit performs industry category preference prediction for the target object based on the basic information and the asset information, when determining the industry category preference information of the target object, is configured to:

14. The apparatus according to claim 13, wherein the preference information determining unit is configured to, when analyzing the feature set of the target object to obtain the industry category preference information of the target object, based on an association relationship between the feature set of the different object and the industry category preference of the different object:

15. The apparatus according to any one of claims 11 to 14, wherein the preference information determining unit performs feature extraction on the basic information and the asset information, and when determining a feature set of the target object, is configured to:

16. The apparatus of claim 15, wherein the asset allocation unit is configured to, when selecting one or more features from the candidate feature set to form the feature set:

17. An electronic device, comprising:

a memory for storing program instructions;

A processor for invoking program instructions stored in the memory and for performing the steps comprised in the method according to any of claims 1-8 in accordance with the obtained program instructions.

18. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program comprising program instructions which, when executed by a computer, cause the computer to perform the method of any of claims 1-8.

19. A computer program product, the computer program product comprising: computer program code which, when run on a computer, causes the computer to perform the method of any of the preceding claims 1-8.