CN119720056A - Customer payment transaction data anomaly recognition system and method based on artificial intelligence - Google Patents
Customer payment transaction data anomaly recognition system and method based on artificial intelligence Download PDFInfo
- Publication number
- CN119720056A CN119720056A CN202510230411.8A CN202510230411A CN119720056A CN 119720056 A CN119720056 A CN 119720056A CN 202510230411 A CN202510230411 A CN 202510230411A CN 119720056 A CN119720056 A CN 119720056A
- Authority
- CN
- China
- Prior art keywords
- payment transaction
- customer payment
- transaction data
- data set
- artificial
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
本发明涉及数据异常识别的技术领域，公开了基于人工智能的客户支付交易数据异常识别系统及方法。本发明首先基于聚类距离对初始客户支付交易数据集合进行数据填补，得到处理后的客户支付交易数据集合，再对处理后的客户支付交易数据集合中少类型客户支付交易数据集合进行数据扩充，得到处理好的客户支付交易数据集合；其次，构建初始客户支付交易混合预测模型，并使用改进的人工兔优化算法优化模型的超参数；最后，结合所述处理好的客户支付交易数据集合，计算客户支付交易异常特征，识别客户支付交易异常数据。本发明通过对客户支付交易数据进行分析处理，实现客户支付交易数据异常识别的目的，方法客观准确。
The present invention relates to the technical field of data anomaly identification, and discloses a customer payment transaction data anomaly identification system and method based on artificial intelligence. The present invention first fills the initial customer payment transaction data set based on cluster distance to obtain a processed customer payment transaction data set, and then expands the data of the small type customer payment transaction data set in the processed customer payment transaction data set to obtain a processed customer payment transaction data set; secondly, constructs an initial customer payment transaction hybrid prediction model, and uses an improved artificial rabbit optimization algorithm to optimize the hyperparameters of the model; finally, in combination with the processed customer payment transaction data set, calculates the abnormal characteristics of the customer payment transaction, and identifies the abnormal data of the customer payment transaction. The present invention achieves the purpose of abnormal identification of customer payment transaction data by analyzing and processing the customer payment transaction data, and the method is objective and accurate.
Description
Technical Field
The invention relates to the technical field of data anomaly identification, in particular to a customer payment transaction data anomaly identification system and method based on artificial intelligence.
Background
The Chinese patent CN111949702B discloses a method, a device and equipment for identifying abnormal transaction data, the method specifically comprises the steps of obtaining characteristic parameters of entity objects in transaction data by using registration information of the entity objects and hot spot information connected with users in the transaction data, analyzing user data in the transaction data by using a data model trained by characteristic labels to obtain the characteristic parameters captured by the users in the entity objects, constructing an abnormal identification model, inputting the characteristic parameters into the abnormal identification model to form multidimensional characteristics, outputting a transaction data identification result, judging association between the transaction data identification result and the entity objects, and regarding corresponding transaction data as abnormal transaction data when the association exists between the transaction data identification result and the entity objects, or else, judging the corresponding transaction data as abnormal transaction data. The method does not preprocess transaction data, and can cause problems such as low recognition speed.
The traditional abnormal identification method of the client payment transaction data has the advantages that due to the characteristics of multiple types of the client payment transaction data, data redundancy and the like, the abnormal client payment transaction data is long in identification time, potential safety hazards are easy to occur, meanwhile, due to the fact that artificial intelligence and other technologies are not used, real-time monitoring of the client payment transaction data is difficult, and network payment fraud transaction behaviors are easy to occur.
Disclosure of Invention
Aiming at the problems in the related art, the invention provides a system and a method for identifying the abnormality of the payment transaction data of a customer based on artificial intelligence, so as to overcome the technical problems existing in the related art.
In order to solve the technical problems, the invention is realized by the following technical scheme:
The invention relates to a customer payment transaction data anomaly identification method based on artificial intelligence, which comprises the following steps:
S1, acquiring a client payment transaction data type to obtain an initial client payment transaction data set, and performing data filling on the initial client payment transaction data set based on a clustering distance to obtain a processed client payment transaction data set;
S2, identifying a few-type customer payment transaction data set and a multi-type customer payment transaction data set in the processed customer payment transaction data set, and performing data expansion on the few-type customer payment transaction data set to obtain a processed customer payment transaction data set;
s3, fusing a convolutional neural network and a long-term and short-term memory neural network, constructing an initial customer payment transaction mixed prediction model, and introducing an improved artificial rabbit optimization algorithm to optimize the super parameters of the customer payment transaction mixed prediction model to obtain optimized super parameters;
S4, the initial customer payment transaction mixed prediction model uses the optimized super parameters to obtain a final customer payment transaction mixed prediction model, and then the final customer payment transaction mixed prediction model is combined with the processed customer payment transaction data set to output a customer payment transaction amount predicted value, and then the customer payment transaction abnormal characteristics are calculated to identify customer payment transaction abnormal data.
The method is superior to the traditional data filling method in filling accuracy and filling speed, the clustering distance is used for replacing similarity, data quality is effectively improved, secondly, data expansion is conducted on fewer types of customer payment transaction data sets in the processed customer payment transaction data sets, interpolation is conducted only on fewer types of data sets, the data sets are expanded, the problem of category unbalance in the data sets is solved, subsequent processing is facilitated, a convolutional neural network and a long-term memory neural network are fused, an initial customer payment transaction mixed prediction model is constructed, super parameters of the customer payment transaction mixed prediction model are optimized by using an improved artificial rabbit optimization algorithm, data characteristics are extracted by the mixed prediction model, model calculation amount and model complexity are reduced, model prediction accuracy is greatly improved by using a long-term memory neural network, improved artificial rabbit optimization algorithm increases searching capacity and a special-Monte-Care flight method compared with the traditional algorithm, the problem that the optimal transaction characteristics are improved, the overall payment transaction performance is easily recognized, and the abnormal value is judged, and the abnormal transaction performance is prevented from being output by the method.
Preferably, the step S1 includes the steps of:
S11, acquiring client payment transaction data types, wherein the client payment transaction data types comprise transaction types, transaction time, payment channels, transaction amounts and the like, setting the number of the client payment transaction data types as m, collecting client payment transaction data under each client payment transaction data type, and quantitatively processing the client payment transaction data to generate an initial client payment transaction data set WhereinRepresenting an mth customer payment transaction data type;
S12, carrying out data filling on the initial customer payment transaction data set based on the clustering distance to obtain a processed customer payment transaction data set, wherein the method comprises the following specific steps of:
S121, judging whether the initial customer payment transaction data set has a missing value according to the fact that the missing value exists or not, adding customer payment transaction data in the customer payment transaction data type into the customer payment transaction missing data set when the missing value exists in the customer payment transaction data type, otherwise adding customer payment transaction data in the customer payment transaction data type into the customer payment transaction complete data set, selecting an initial clustering center point in the customer payment transaction complete data set, traversing the customer payment transaction complete data set, calculating to obtain customer payment transaction data closest to the initial clustering center point in the customer payment transaction complete data set, merging the customer payment transaction data into a clustering cluster of the initial clustering center point, calculating the square sum of the distances of customer payment transaction data in the clustering cluster of the initial clustering center point, and taking the sum of the distances as a second clustering center point;
S122, taking the final clustering center point as a filling value, filling the missing value in the customer payment transaction missing data set for the first time to obtain a preliminary filled customer payment transaction data set, calculating the distance between customer payment transaction data and the filling value in the preliminary filled customer payment transaction data set, marking the distance as a clustering distance, setting a distance threshold, not filling the preliminary filled customer payment transaction data set for the second time when the distance between the filling value and the minimum value of the clustering distance is smaller than the distance threshold, otherwise replacing the filling value by using the minimum value of the clustering distance, filling the preliminary filled customer payment transaction data set for the second time until all customer payment transaction data types are filled, and obtaining the processed customer payment transaction data set.
According to the method, the clustering center of the complete data set is found, the clustering distance is calculated, the similarity is replaced by the clustering distance, the data quality is effectively improved, the missing data set is filled according to the clustering distance, and the filling accuracy and the filling speed are superior to those of the traditional data filling method.
Preferably, the step S2 includes the steps of:
S21, setting a data number threshold, when the number of the client payment transaction data types in the processed client payment transaction data set is smaller than the data number threshold, marking the corresponding client payment transaction data types as few-type client payment transaction data sets, otherwise marking the corresponding client payment transaction data types as multi-type client payment transaction data sets;
When the nearest neighbor data in the first nearest neighbor data set is larger than the payment transaction data of the undetermined customer, regarding the payment transaction data of the undetermined customer as noise data, and reselecting other payment transaction data of the customer; when the nearest neighbor data in the first nearest neighbor data set is smaller than or equal to half of the pending customer payment transaction data and larger than or equal to half of the pending customer payment transaction data, the pending customer payment transaction data are added into the pending customer payment transaction data set;
S22, calculating the nearest neighbor data of the pending customer payment transaction data in the less-type customer payment transaction data set in the pending customer payment transaction data set to obtain a second nearest neighbor data set WhereinSelecting pending customer payment transaction data from the set of pending customer payment transaction dataCalculating pending customer payment transaction dataThe distance to the second nearest neighbor data set is noted asWhereinRepresenting pending customer payment transaction dataDistance to the c-th nearest neighbor data, at which time the transaction data is paid for by the customer to be determinedAnd (3) data expansion, wherein the calculation formula is as follows:
;
Wherein, Representing pending customer payment transaction dataIs used for the expansion of the data,Representing pending customer payment transaction dataTo the firstThe distance of the nearest neighbor data,;
And sequentially carrying out data expansion on all pending customer payment transaction data in the pending customer payment transaction data set to obtain an expanded customer payment transaction data set, and generating a processed customer payment transaction data set by combining the multi-type customer payment transaction data set.
According to the invention, the data expansion is carried out on the few types of customer payment transaction data sets in the processed customer payment transaction data sets, interpolation is carried out only on the few types of data sets, the data sets are expanded, the operation time consumption is reduced, the problem of class unbalance in the data sets is solved, and the subsequent processing is convenient.
Preferably, the step S3 includes the steps of:
s31, fusing a convolutional neural network and a long-term memory neural network, and constructing to obtain an initial customer payment transaction mixed prediction model, wherein the method comprises the following specific steps of:
S311, giving a time sequence to the processed customer payment transaction data set according to the transaction time to obtain a customer payment transaction data time sequence; setting a convolutional neural network comprising an input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer, adding an attention mechanism after the input layer, and adding a long-term memory neural network between the pooling layer and the full-connection layer;
S312, after data samples are input from an input layer of the initial customer payment transaction mixed prediction model, data screening is carried out through an attention mechanism, the data samples are input into a convolution layer and a pooling layer for data feature extraction, then a long-period memory neural network is used for obtaining space-time correlation by combining the customer payment transaction data time sequence, a predicted value of the data samples is output, and the initial customer payment transaction mixed prediction model is constructed;
S32, acquiring a last-year customer payment transaction data set, carrying out data filling and data expansion on the last-year customer payment transaction data set to obtain a customer payment transaction data sample set, endowing a time sequence, inputting the customer payment transaction data sample set into an initial customer payment transaction mixed prediction model according to the time sequence for training, using an improved artificial rabbit optimization algorithm to optimize super parameters in the initial customer payment transaction mixed prediction model to obtain optimized super parameters, and specifically, the method comprises the following steps of:
S321, setting an artificial rabbit population in which the number of the artificial rabbit population is i, the individual dimension of the artificial rabbit in the artificial rabbit population is j, the individual position of the artificial rabbit in the artificial rabbit population represents a super parameter, initializing the artificial rabbit population to determine the initial position and the initial fitness function value of the individual of the artificial rabbit, setting the current iteration number as d in the roundabout feeding process of the artificial rabbit population, and recording the individual position of the e-th artificial rabbit as d in the d iteration The d th iterationThe individual positions of only artificial rabbits are recorded as,Representing random numbers between [0,1] intervals, the standard normal distribution coefficient isSetting running factor, and replacing running factor with Lewy flight coefficient of,AndRepresenting random numbers between (0, 1) intervals, then the Lewy flight functionThe individual position of the e artificial rabbit at the (d+1) th iterationThe calculation formula is as follows:
;
Wherein, Representing a rounding function;
s322, in the random hiding process of the artificial rabbit population, the artificial rabbit individuals select caves to hide, and the hiding parameters are set as ,Representing an integer of 0 or 1, then the h cave location of the e-th artificial rabbit individualThe cave position is used for the individual position of the artificial rabbitUpdate at this timeIntroducing Monte Carlo method to update the individual position of the artificial rabbit againCalculating the average position of the individual artificial rabbits in the artificial rabbit population at the momentThe control parameters are,Representing a random number between the [0,1] intervals,Representing random numbers between (0, 1) intervals whenIn the time-course of which the first and second contact surfaces,When (when)In the time-course of which the first and second contact surfaces,;
S323, calculating an fitness function value corresponding to the individual position of the e-th artificial rabbit in the d-th iterationFitness function value corresponding to individual position of e-th artificial rabbit at d+1st iterationCorresponding to the super parameters in the initial customer payment transaction mixed prediction model at the d iteration and the (d+1) iteration respectively, comparing fitness function valuesFitness function valueSetting the maximum iteration number, stopping iteration when the current iteration number reaches the maximum iteration number, and obtaining a global optimal position, wherein the global optimal position corresponds to the optimized super-parameter.
The method comprises the steps of establishing an initial customer payment transaction mixed prediction model by fusing a convolutional neural network and a long-short-term memory neural network, optimizing super parameters of the customer payment transaction mixed prediction model by using an improved artificial rabbit optimization algorithm, extracting data features by the mixed prediction model by using the convolutional neural network, reducing the calculated amount and complexity of the model, solving the gradient expansion disappearance problem by the long-short-term memory neural network, greatly improving the model prediction precision, and improving the search capability and the local development capability of the algorithm by introducing the Lewy flight and Monte Carlo method by the improved artificial rabbit optimization algorithm compared with the traditional algorithm, so that the trouble of local optimal solution is avoided, and the overall performance is improved.
Preferably, the step S4 includes the steps of:
s41, the initial customer payment transaction mixed prediction model uses optimized super parameters until the initial customer payment transaction mixed prediction model converges to obtain a final customer payment transaction mixed prediction model, the processed customer payment transaction data set is used as a time sequence according to transaction time to obtain a final customer payment transaction data set, a time sequence step is set, the final customer payment transaction data set is subjected to sectional processing by using the time sequence step, and then the final customer payment transaction data set is input into the final customer payment transaction mixed prediction model to output a customer payment transaction amount predicted value;
S42, outputting predicted value of the customer payment transaction amount every time sequence step length to obtain a predicted value set of the customer payment transaction amount, calculating time intervals of the predicted value of the customer payment transaction amount, recording the time intervals as the time intervals of the customer payment transaction, and calculating time periods The average value of the predicted value of the client payment transaction amount in the range is recorded as the average client payment transaction amount, and the time period is calculatedThe average value of the predicted value quantity of the client payment transaction amount in the range is recorded as the client payment transaction frequency, and the client payment transaction time interval, the average client payment transaction amount and the client payment transaction frequency are combined to be used as abnormal characteristics of the client payment transaction;
Respectively setting a first threshold value A second threshold valueAnd a third threshold valueWhen the time interval of the client payment transaction is smaller than a first threshold value or the average client payment transaction amount is larger than a second threshold value or the frequency of the client payment transaction is larger than a third threshold value, the client payment transaction data is abnormal, and the client feeds back abnormal client payment transaction data to further process and survey so as to complete abnormal identification of the client payment transaction data.
The embodiment also discloses a system of the client payment transaction data anomaly identification method based on artificial intelligence, which specifically comprises a payment transaction data filling module, a payment transaction data expansion module, a super-parameter optimization module and a data anomaly identification module;
the payment transaction data filling module is used for filling data into the initial customer payment transaction data set based on the clustering distance;
The payment transaction data expansion module is used for carrying out data expansion on a few-type customer payment transaction data set;
the super-parameter optimization module is used for optimizing super-parameters of the customer payment transaction hybrid prediction model by using an improved artificial rabbit optimization algorithm;
The data anomaly identification module is used for identifying abnormal data of the customer payment transaction after calculating abnormal characteristics of the customer payment transaction.
The invention has the following beneficial effects:
1. According to the method, the clustering center of the complete data set is found, the clustering distance is calculated, the similarity is replaced by the clustering distance, the data quality is effectively improved, the missing data set is filled according to the clustering distance, and the filling accuracy and the filling speed are superior to those of the traditional data filling method.
2. According to the invention, the data expansion is carried out on the few types of customer payment transaction data sets in the processed customer payment transaction data sets, interpolation is carried out only on the few types of data sets, the data sets are expanded, the operation time consumption is reduced, the problem of class unbalance in the data sets is solved, and the subsequent processing is convenient.
3. According to the invention, the convolutional neural network and the long-short-period memory neural network are fused to construct an initial customer payment transaction hybrid prediction model, the data characteristics are extracted by the convolutional neural network through the hybrid prediction model, the calculated amount and complexity of the model are reduced, the gradient expansion disappearance problem is processed by the long-short-period memory neural network, and the model prediction precision is greatly improved.
4. According to the invention, the super parameters of the customer payment transaction mixed prediction model are optimized by using the improved artificial rabbit optimization algorithm, and the improved artificial rabbit optimization algorithm increases the searching capability and the local development capability of the algorithm compared with the traditional algorithm by introducing the Lewy flight and Monte Carlo methods, so that the trouble of local optimal solution is avoided, and the overall performance is improved.
Of course, it is not necessary for any one product to practice the invention to achieve all of the advantages set forth above at the same time.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings that are needed for the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the invention, and that it is also possible for a person skilled in the art to obtain the drawings from these drawings without inventive effort.
Fig. 1 is a schematic flow chart of the abnormal identification system of the customer payment transaction data based on artificial intelligence in the abnormal identification of the customer payment transaction data.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments that can be obtained by a person of ordinary skill in the art without making any inventive effort are within the scope of the present invention.
Embodiment 1 referring to fig. 1, the present embodiment discloses an artificial intelligence based method for identifying anomalies in customer payment transaction data, which specifically includes the following steps:
S1, acquiring a client payment transaction data type to obtain an initial client payment transaction data set, and performing data filling on the initial client payment transaction data set based on a clustering distance to obtain a processed client payment transaction data set;
the step S1 comprises the following steps:
S11, acquiring client payment transaction data types, wherein the client payment transaction data types comprise transaction types, transaction time, payment channels, transaction amounts and the like, setting the number of the client payment transaction data types as m, collecting client payment transaction data under each client payment transaction data type, and quantitatively processing the client payment transaction data to generate an initial client payment transaction data set WhereinRepresenting an mth customer payment transaction data type;
S12, carrying out data filling on the initial customer payment transaction data set based on the clustering distance to obtain a processed customer payment transaction data set, wherein the method comprises the following specific steps of:
S121, judging whether the initial customer payment transaction data set has a missing value according to the fact that the missing value exists or not, adding customer payment transaction data in the customer payment transaction data type into the customer payment transaction missing data set when the missing value exists in the customer payment transaction data type, otherwise adding customer payment transaction data in the customer payment transaction data type into the customer payment transaction complete data set, selecting an initial clustering center point in the customer payment transaction complete data set, traversing the customer payment transaction complete data set, calculating to obtain customer payment transaction data closest to the initial clustering center point in the customer payment transaction complete data set, merging the customer payment transaction data into a clustering cluster of the initial clustering center point, calculating the square sum of the distances of customer payment transaction data in the clustering cluster of the initial clustering center point, and taking the sum of the distances as a second clustering center point;
s122, taking the final clustering center point as a filling value, filling the missing value in the customer payment transaction missing data set for the first time to obtain a preliminary filled customer payment transaction data set, calculating the distance between customer payment transaction data and the filling value in the preliminary filled customer payment transaction data set, marking the distance as a clustering distance, setting a distance threshold, and not filling the preliminary filled customer payment transaction data set for the second time when the distance between the filling value and the minimum value of the clustering distance is smaller than the distance threshold, otherwise, replacing the filling value by using the minimum value of the clustering distance, and filling the preliminary filled customer payment transaction data set for the second time until all the customer payment transaction data types are filled, thereby obtaining the processed customer payment transaction data set;
S2, identifying a few-type customer payment transaction data set and a multi-type customer payment transaction data set in the processed customer payment transaction data set, and performing data expansion on the few-type customer payment transaction data set to obtain a processed customer payment transaction data set;
The step S2 comprises the following steps:
S21, setting a data number threshold, when the number of the client payment transaction data types in the processed client payment transaction data set is smaller than the data number threshold, marking the corresponding client payment transaction data types as few-type client payment transaction data sets, otherwise marking the corresponding client payment transaction data types as multi-type client payment transaction data sets;
When the nearest neighbor data in the first nearest neighbor data set is larger than the payment transaction data of the undetermined customer, regarding the payment transaction data of the undetermined customer as noise data, and reselecting other payment transaction data of the customer; when the nearest neighbor data in the first nearest neighbor data set is smaller than or equal to half of the pending customer payment transaction data and larger than or equal to half of the pending customer payment transaction data, the pending customer payment transaction data are added into the pending customer payment transaction data set;
S22, calculating the nearest neighbor data of the pending customer payment transaction data in the less-type customer payment transaction data set in the pending customer payment transaction data set to obtain a second nearest neighbor data set WhereinSelecting pending customer payment transaction data from the set of pending customer payment transaction dataCalculating pending customer payment transaction dataThe distance to the second nearest neighbor data set is noted asWhereinRepresenting pending customer payment transaction dataDistance to the c-th nearest neighbor data, at which time the transaction data is paid for by the customer to be determinedAnd (3) data expansion, wherein the calculation formula is as follows:
;
Wherein, Representing pending customer payment transaction dataIs used for the expansion of the data,Representing pending customer payment transaction dataTo the firstThe distance of the nearest neighbor data,;
Sequentially carrying out data expansion on all pending customer payment transaction data in the pending customer payment transaction data set to obtain an expanded customer payment transaction data set, and generating a processed customer payment transaction data set by combining the multi-type customer payment transaction data set;
s3, fusing a convolutional neural network and a long-term and short-term memory neural network, constructing an initial customer payment transaction mixed prediction model, and introducing an improved artificial rabbit optimization algorithm to optimize the super parameters of the customer payment transaction mixed prediction model to obtain optimized super parameters;
the step S3 comprises the following steps:
s31, fusing a convolutional neural network and a long-term memory neural network, and constructing to obtain an initial customer payment transaction mixed prediction model, wherein the method comprises the following specific steps of:
S311, giving a time sequence to the processed customer payment transaction data set according to the transaction time to obtain a customer payment transaction data time sequence; setting a convolutional neural network comprising an input layer, a convolutional layer, a pooling layer, a full-connection layer and an output layer, adding an attention mechanism after the input layer, and adding a long-term memory neural network between the pooling layer and the full-connection layer;
S312, after data samples are input from an input layer of the initial customer payment transaction mixed prediction model, data screening is carried out through an attention mechanism, the data samples are input into a convolution layer and a pooling layer for data feature extraction, then a long-period memory neural network is used for obtaining space-time correlation by combining the customer payment transaction data time sequence, a predicted value of the data samples is output, and the initial customer payment transaction mixed prediction model is constructed;
S32, acquiring a last-year customer payment transaction data set, carrying out data filling and data expansion on the last-year customer payment transaction data set to obtain a customer payment transaction data sample set, endowing a time sequence, inputting the customer payment transaction data sample set into an initial customer payment transaction mixed prediction model according to the time sequence for training, using an improved artificial rabbit optimization algorithm to optimize super parameters in the initial customer payment transaction mixed prediction model to obtain optimized super parameters, and specifically, the method comprises the following steps of:
S321, setting an artificial rabbit population in which the number of the artificial rabbit population is i, the individual dimension of the artificial rabbit in the artificial rabbit population is j, the individual position of the artificial rabbit in the artificial rabbit population represents a super parameter, initializing the artificial rabbit population to determine the initial position and the initial fitness function value of the individual of the artificial rabbit, setting the current iteration number as d in the roundabout feeding process of the artificial rabbit population, and recording the individual position of the e-th artificial rabbit as d in the d iteration The d th iterationThe individual positions of only artificial rabbits are recorded as,Representing random numbers between [0,1] intervals, the standard normal distribution coefficient isSetting running factor, and replacing running factor with Lewy flight coefficient of,AndRepresenting random numbers between (0, 1) intervals, then the Lewy flight functionThe individual position of the e artificial rabbit at the (d+1) th iterationThe calculation formula is as follows:
;
Wherein, Representing a rounding function;
s322, in the random hiding process of the artificial rabbit population, the artificial rabbit individuals select caves to hide, and the hiding parameters are set as ,Representing an integer of 0 or 1, then the h cave location of the e-th artificial rabbit individualThe cave position is used for the individual position of the artificial rabbitUpdate at this timeIntroducing Monte Carlo method to update the individual position of the artificial rabbit againCalculating the average position of the individual artificial rabbits in the artificial rabbit population at the momentThe control parameters are,Representing a random number between the [0,1] intervals,Representing random numbers between (0, 1) intervals whenIn the time-course of which the first and second contact surfaces,When (when)In the time-course of which the first and second contact surfaces,;
S323, calculating an fitness function value corresponding to the individual position of the e-th artificial rabbit in the d-th iterationFitness function value corresponding to individual position of e-th artificial rabbit at d+1st iterationCorresponding to the super parameters in the initial customer payment transaction mixed prediction model at the d iteration and the (d+1) iteration respectively, comparing fitness function valuesFitness function valueSetting the maximum iteration times, stopping iteration when the current iteration times reach the maximum iteration times, and obtaining a global optimal position, wherein the global optimal position corresponds to the optimized super-parameter;
S4, the initial customer payment transaction mixed prediction model uses the optimized super parameters to obtain a final customer payment transaction mixed prediction model, and then the final customer payment transaction mixed prediction model is combined with the processed customer payment transaction data set to output a customer payment transaction amount predicted value, and then the customer payment transaction abnormal characteristics are calculated to identify customer payment transaction abnormal data;
The step S4 comprises the following steps:
s41, the initial customer payment transaction mixed prediction model uses optimized super parameters until the initial customer payment transaction mixed prediction model converges to obtain a final customer payment transaction mixed prediction model, the processed customer payment transaction data set is used as a time sequence according to transaction time to obtain a final customer payment transaction data set, a time sequence step is set, the final customer payment transaction data set is subjected to sectional processing by using the time sequence step, and then the final customer payment transaction data set is input into the final customer payment transaction mixed prediction model to output a customer payment transaction amount predicted value;
S42, outputting predicted value of the customer payment transaction amount every time sequence step length to obtain a predicted value set of the customer payment transaction amount, calculating time intervals of the predicted value of the customer payment transaction amount, recording the time intervals as the time intervals of the customer payment transaction, and calculating time periods The average value of the predicted value of the client payment transaction amount in the range is recorded as the average client payment transaction amount, and the time period is calculatedThe average value of the predicted value quantity of the client payment transaction amount in the range is recorded as the client payment transaction frequency, and the client payment transaction time interval, the average client payment transaction amount and the client payment transaction frequency are combined to be used as abnormal characteristics of the client payment transaction;
Respectively setting a first threshold value A second threshold valueAnd a third threshold valueWhen the time interval of the client payment transaction is smaller than a first threshold value or the average client payment transaction amount is larger than a second threshold value or the frequency of the client payment transaction is larger than a third threshold value, the client payment transaction data is abnormal, and the client feeds back abnormal client payment transaction data to further process and survey so as to complete abnormal identification of the client payment transaction data.
The embodiment 2 also discloses a system of the client payment transaction data anomaly identification method based on artificial intelligence, which specifically comprises a payment transaction data filling module, a payment transaction data expansion module, a super-parameter optimization module and a data anomaly identification module;
the payment transaction data filling module is used for filling data into the initial customer payment transaction data set based on the clustering distance;
The payment transaction data expansion module is used for carrying out data expansion on a few-type customer payment transaction data set;
the super-parameter optimization module is used for optimizing super-parameters of the customer payment transaction hybrid prediction model by using an improved artificial rabbit optimization algorithm;
The data anomaly identification module is used for identifying abnormal data of the customer payment transaction after calculating abnormal characteristics of the customer payment transaction.
In the description of the present specification, the descriptions of the terms "one embodiment," "example," "specific example," and the like, mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The above disclosed preferred embodiments of the invention are merely intended to help illustrate the invention. The preferred embodiments are not exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention.
Claims (9)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510230411.8A CN119720056A (en) | 2025-02-28 | 2025-02-28 | Customer payment transaction data anomaly recognition system and method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510230411.8A CN119720056A (en) | 2025-02-28 | 2025-02-28 | Customer payment transaction data anomaly recognition system and method based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119720056A true CN119720056A (en) | 2025-03-28 |
Family
ID=95088395
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510230411.8A Pending CN119720056A (en) | 2025-02-28 | 2025-02-28 | Customer payment transaction data anomaly recognition system and method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119720056A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8140399B1 (en) * | 2003-10-24 | 2012-03-20 | Sachin Goel | System for concurrent optimization of business economics and customer value |
| CN117036634A (en) * | 2023-10-08 | 2023-11-10 | 青岛星邦光电科技有限责任公司 | Automatic construction method for three-dimensional scene of smart city |
| CN118152992A (en) * | 2024-01-25 | 2024-06-07 | 西南石油大学 | Pipeline corrosion prediction method based on machine learning |
| CN118195780A (en) * | 2024-04-11 | 2024-06-14 | 中国工商银行股份有限公司 | Transaction monitoring method, device, apparatus, medium and program product |
| CN118797275A (en) * | 2024-09-11 | 2024-10-18 | 湖北工业大学 | Power load forecasting method and system based on improved grey wolf optimization algorithm |
-
2025
- 2025-02-28 CN CN202510230411.8A patent/CN119720056A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8140399B1 (en) * | 2003-10-24 | 2012-03-20 | Sachin Goel | System for concurrent optimization of business economics and customer value |
| CN117036634A (en) * | 2023-10-08 | 2023-11-10 | 青岛星邦光电科技有限责任公司 | Automatic construction method for three-dimensional scene of smart city |
| CN118152992A (en) * | 2024-01-25 | 2024-06-07 | 西南石油大学 | Pipeline corrosion prediction method based on machine learning |
| CN118195780A (en) * | 2024-04-11 | 2024-06-14 | 中国工商银行股份有限公司 | Transaction monitoring method, device, apparatus, medium and program product |
| CN118797275A (en) * | 2024-09-11 | 2024-10-18 | 湖北工业大学 | Power load forecasting method and system based on improved grey wolf optimization algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113378632B (en) | An unsupervised domain adaptation person re-identification method based on pseudo-label optimization | |
| CN111612041B (en) | Abnormal user identification method and device, storage medium and electronic equipment | |
| CN111401599B (en) | Water level prediction method based on similarity search and LSTM neural network | |
| CN113297936B (en) | Volleyball group behavior identification method based on local graph convolution network | |
| CN112784173B (en) | Recommendation system scoring prediction method based on self-attention confrontation neural network | |
| CN112700324A (en) | User loan default prediction method based on combination of Catboost and restricted Boltzmann machine | |
| Abdul-Rahman et al. | Advanced machine learning algorithms for house price prediction: case study in Kuala Lumpur | |
| CN110990718A (en) | Social network model building module of company image improving system | |
| CN114819392B (en) | A method for predicting electricity consumption based on user clustering and extended data | |
| CN117236666B (en) | Emergency material demand analysis method and system | |
| CN117539920B (en) | Data query method and system based on real estate transaction multidimensional data | |
| CN118964933A (en) | User financial management behavior analysis method based on artificial intelligence and financial big data | |
| CN114997366A (en) | Protein structure model quality evaluation method based on graph neural network | |
| CN118196662A (en) | A method for establishing a coal mine digital twin model based on Bayesian network algorithm | |
| CN119474552A (en) | A cultural and tourism content recommendation system that analyzes preferences for cultural and tourism attractions | |
| CN119323452A (en) | Cloud computing-based customer data integrated management system and method | |
| CN120182852A (en) | Cross-modal remote sensing target detection method and system based on spatial consistency constraint and deep feature alignment | |
| Ragapriya et al. | Machine learning based house price prediction using modified extreme boosting | |
| CN112650949A (en) | Regional POI (Point of interest) demand identification method based on multi-source feature fusion collaborative filtering | |
| CN120277571A (en) | Construction method of multidimensional time sequence anomaly detection system | |
| CN115905893A (en) | Resource numerical value prediction method, device, computer equipment and storage medium | |
| CN119398787A (en) | A card transaction warning method and device based on AI recognition | |
| CN119358947A (en) | A land resource management information platform based on cloud computing | |
| CN119249236A (en) | A user profile prediction method based on sparse autoencoder and K-means | |
| CN119720056A (en) | Customer payment transaction data anomaly recognition system and method based on artificial intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |