CN113888252B - Recommendation method based on user's food safety rating and food similarity - Google Patents
Recommendation method based on user's food safety rating and food similarity Download PDFInfo
- Publication number
- CN113888252B CN113888252B CN202111003703.6A CN202111003703A CN113888252B CN 113888252 B CN113888252 B CN 113888252B CN 202111003703 A CN202111003703 A CN 202111003703A CN 113888252 B CN113888252 B CN 113888252B
- Authority
- CN
- China
- Prior art keywords
- food
- scoring
- user
- data
- matrix
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Recommending goods or services
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0623—Electronic shopping [e-shopping] by investigating goods or services
- G06Q30/0625—Electronic shopping [e-shopping] by investigating goods or services by formulating product or service queries, e.g. using keywords or predefined options
- G06Q30/0629—Electronic shopping [e-shopping] by investigating goods or services by formulating product or service queries, e.g. using keywords or predefined options by pre-processing results, e.g. ranking or ordering results
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
本发明公开了一种基于用户对食品安全等级评分值和食品相似度的推荐方法，包括步骤：1)获取用户对食品的评分数据；2)计算出每一条评分数据的评分权重。3)将评分数据和评分权重输入到机器学习模型中进行参数训练。4)参数训练完成后，得到一个食品相似度矩阵，最终通过用户的评分数据和食品相似度矩阵计算并生成用户的食品推荐列表，实现将食品推荐给用户。本发明在训练出一个机器学习模型的同时结合使用了基于邻域的协同过滤方法，从用户对食品的评分数据中学习出食品的相似度矩阵，考虑评分数据时间顺序的同时将稀疏性引入相似度矩阵，使其能够有效地生成推荐。
The present invention discloses a recommendation method based on user's food safety grade rating and food similarity, comprising the steps of: 1) obtaining user's food rating data; 2) calculating the rating weight of each rating data. 3) inputting the rating data and the rating weight into a machine learning model for parameter training. 4) After the parameter training is completed, a food similarity matrix is obtained, and finally the user's food recommendation list is calculated and generated through the user's rating data and the food similarity matrix, so as to recommend food to the user. The present invention combines the use of a neighborhood-based collaborative filtering method while training a machine learning model, learns the food similarity matrix from the user's food rating data, introduces sparsity into the similarity matrix while considering the time sequence of the rating data, so that it can effectively generate recommendations.
Description
Technical Field
The invention relates to the technical field of food recommendation, in particular to a recommendation method based on food safety grade grading values and food similarity of users.
Background
The advent and rapid growth of electronic commerce has greatly changed the traditional concept of purchasing food by people, which has made online transactions easier by providing massive amounts of food and detailed food information. However, as the number of food products meeting the needs of users has increased dramatically, it has become a problem how to effectively and efficiently assist users in identifying food products that best fit their personal preferences, especially given the user's safety rating score value data for the food products. Generating a list of recommended items for the user by means of a recommendation algorithm is a widely used working objective of TopN recommendation systems. In recent years, various TopN recommendation algorithms have been developed. These algorithms can be divided into two categories, neighborhood-based collaborative filtering methods and model-based collaborative filtering methods. The neighborhood based approach can directly calculate the similarity of the user or the item only by the user's behavioral data, where the item neighborhood based approach can generate recommendations very quickly, which is simple to implement, but it does so at the expense of recommendation effectiveness. Model-based methods require learning of user or item feature data from user behavior data, particularly potential factor model-based methods, while the machine learning model may generate better recommendation quality, at the same time, higher time costs may be incurred in the model's parameter learning. On the other hand, in either the neighborhood-based method or the model-based method, the specific scoring value of the item by the user is directly used in the selection of the data sample, and various factors which may influence the scoring of the user are ignored. The user's previous scoring of some items is highly likely to mislead the scoring of the current item and thus affect the authenticity and reliability of the scoring. In this case, the similarity between the items is affected by the order of the scoring of the users, so that the subjective scores of the users cannot objectively reflect the similarity between the items.
Disclosure of Invention
The invention aims to overcome the defects and shortcomings of the prior art, and provides a recommendation method based on food safety grade grading value and food similarity of users, which breaks through the problems of low recommendation effect based on a neighborhood method and low recommendation efficiency based on a model method, the similarity matrix of the food is learned from the safety grade grading value data of the food of the user, and sparsity is introduced into the similarity matrix while the time sequence of the grading data is considered, so that the recommendation can be effectively generated.
In order to achieve the above purpose, the technical scheme provided by the invention is that the recommendation method based on the food safety grade grading value and the food similarity of the user comprises the following steps:
1) The method comprises the steps of obtaining grading data of a user on food, wherein each piece of grading data comprises a user ID, food ID, a security grade grading value and grading time;
2) For scoring data of the same user ID, calculating the scoring weight of each piece of scoring data according to the sequence before and after the scoring time and the security grade scoring value;
3) Inputting the user ID of the grading data, the food ID, the safety grade grading value and the grading weight into a machine learning model for parameter training, obtaining a food similarity matrix after the parameter training is completed, and finally calculating and generating a food recommendation list of the user through the grading data of the user and the food similarity matrix to realize food recommendation to the user.
Further, in step 1), the scoring data of the user on the food is user history scoring data for inputting a machine learning model for training, the scoring data comprises a user ID, a food ID, a security grade scoring value and scoring time, wherein the user ID is a unique identifier of the user, different user IDs represent different users, the food ID is a unique identifier of the food, different food IDs represent different foods, the security grade scoring value is the security grade scoring value of the food represented by the food ID by the user represented by the user ID, and the scoring time is the time for scoring the food security grade by the user.
Further, in step 2), the scoring data of different users are grouped according to the user ID, the scoring data of the same user ID are arranged in sequence according to the scoring time, and the scoring weight value of each piece of the scoring data after the sorting is processed according to the following three conditions that ① if the security grade scoring value of the previous piece of the scoring data is the highest scoring value, the scoring weight of the current scoring data is 1.2, ② if the security grade scoring value of the previous piece of the scoring data is the lowest scoring value, the scoring weight of the current scoring data is 0.8, and the scoring weights of the first piece of the scoring data and other pieces of the scoring data are 1.
Further, in step 3), the scoring data of all the users are converted into a scoring matrix, each row of the scoring matrix represents a user, each column of the scoring matrix represents a food, each value of the matrix represents the security grade scoring value of the row of the user on the column of food, the scoring matrix is input into a machine learning model for parameter training, and the matrix form objective function of the machine learning model is as follows:
Wherein R is a scoring matrix of m rows and n columns, m represents the number of users, n represents the number of foods, each row of the scoring matrix represents a user, each column represents a food, W is a training-obtained n-row and n-column food similarity matrix, n represents the number of foods, and each row or each column represents a food, each value in the food similarity matrix represents the similarity of the row of foods and the column of foods, the value is greater than or equal to zero, and meanwhile, the value on the diagonal of the food similarity matrix is zero, the L1 norm of the matrix is equal to the L 1, the L F is the Frobenius norm of the matrix, and the constants beta and lambda are regularization parameters;
Since each column in the food similarity matrix W is independent, each column of the food similarity matrix can be trained in parallel, and the matrix-form objective function can be decomposed into:
Wherein R j is the j-th column of the scoring matrix R, W j is the j-th column of the food similarity matrix W, I I.I. 2 is the L2 norm of the vector, I I.I. 1 is the L1 norm of the vector, constants beta and lambda are regularization parameters, in order to reduce the training time of model parameters, the safety grade scoring value of each piece of scoring data is multiplied by the corresponding scoring weight, a neighborhood-based collaborative filtering method is used for calculating the column with high similarity between R j and the scoring matrix, and only the column with high similarity between R j is selected to fit an objective function in the minimization of the objective function;
the error value formula for calculating each piece of scoring data in the objective function is:
error=weight(rui-predict)
Wherein r ui represents the security grade score value of the ith row and the ith column of the scoring matrix, predict represents the predicted security grade score value, the predicted security grade score value is obtained by calculating the inner product of the ith row of the scoring matrix and the ith column of the food similarity matrix, and the weight error value of each piece of scoring data is as follows under the condition of considering the scoring weight of the scoring data:
WeightError=weight*error
The machine learning model minimizes the weight error value of all the scoring data;
After the scoring data of all users are input into the machine learning model for parameter training, a food similarity matrix W is obtained, and the predicted safety grade scoring value of the un-scored food i of the user u pair is calculated as follows:
PredictRateui＝RuWi
the method comprises the steps of calculating predicted safety grade grading values of all unscored foods by a user, sorting the predicted safety grade grading values, selecting a plurality of foods with highest grading values for recommending to the user, and completing the recommendation process, wherein R u represents a vector of a u th row in a grading matrix, namely grading data of the user u, W i represents a vector of an i th column in a food similarity matrix, namely similarity values of the food i and other foods.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. According to the invention, a machine learning model is trained, and a collaborative filtering method based on a neighborhood is combined, so that the model training time is shortened, and the recommending effect is improved.
2. The invention uses the L1 norm, the L2 norm and the Frobenius norm, ensures the sparsity of the matrix obtained by model training, and prevents the over-fitting phenomenon.
3. According to the method, the weight value of the scoring data is calculated based on the time sequence of the scoring data, so that the authenticity and objectivity of the food similarity matrix obtained by model training are improved, and the recommending effect is improved.
Drawings
FIG. 1 is a schematic logic flow diagram of the method of the present invention.
FIG. 2 is a flowchart of the present invention for calculating scoring weights.
Detailed Description
The invention will be further illustrated with reference to specific examples.
As shown in fig. 1, the recommendation method based on the food safety grade grading value and the food similarity of the user provided in the present embodiment uses a collaborative filtering method based on a neighborhood in combination with training a machine learning model, learns a similarity matrix of food from grading data of food of the user, and introduces sparsity into the similarity matrix while considering the time sequence of the grading data, so as to effectively generate recommendation, and includes the following steps:
1) And obtaining grading data of the user on the food, namely, historical grading data of the user for inputting a machine learning model for training, wherein the grading data comprises a user ID, a food ID, a security grade grading value and grading time. The food safety grade grading method comprises the steps of enabling a user ID to be a unique identifier of the user, enabling different user IDs to represent different users, enabling food IDs to be unique identifiers of foods, enabling different food IDs to represent different foods, enabling a safety grade grading value to be the safety grade grading value of the user represented by the user ID to the food represented by the food ID, and enabling grading time to be the time for the user to grade the food.
2) The scoring data for different users is grouped according to user ID. The scoring data of the same user ID are arranged in sequence according to the scoring time, and for each piece of the scoring data after the ranking, as shown in FIG. 2, the scoring weight value is treated in the following three cases that ① if the security grade scoring value of the previous piece of scoring data is the highest scoring value, the scoring weight of the current scoring data is 1.2, ② if the security grade scoring value of the previous piece of scoring data is the lowest scoring value, the scoring weight of the current scoring data is 0.8, and the scoring weights of the first piece of scoring data and other data are 1.
3) The scoring data for all users is converted into a scoring matrix, each row of the scoring matrix representing a user, each column representing a food item, and each value of the matrix representing the security level scoring value of the row of users for the column of food items. Inputting the scoring matrix into a machine learning model for parameter training, wherein the matrix form objective function of the machine learning model is as follows:
Wherein R is a scoring matrix of m rows and n columns, m represents the number of users, n represents the number of foods, each row of the scoring matrix represents a user, each column represents a food, W is a training-obtained food similarity matrix of n rows and n columns, n represents the number of foods, each row or each column represents a food, each value in the food similarity matrix represents the similarity of the food of the row and the food of the column, the value is greater than or equal to zero, meanwhile, the value on the diagonal of the food similarity matrix is zero, I· 1 is the L1 norm of the matrix, I·I F is the Frobenius norm of the matrix, and constants beta and lambda are regularization parameters.
Since each column in the food similarity matrix W is independent, each column of the food similarity matrix may be trained in parallel, and the matrix-form objective function may be decomposed into:
Wherein R j is the j-th column of the scoring matrix R, W j is the j-th column of the food similarity matrix W, I I.I 2 is the L2 norm of the vector, I I.I 1 is the L1 norm of the vector, and the constants beta and lambda are regularization parameters. To reduce the time for model parameter training, a neighborhood-based collaborative filtering method is used in advance to calculate columns in the scoring matrix that have high similarity to R j, and only columns with high similarity to R j are selected to fit the objective function in minimizing the objective function.
The error value formula for calculating each piece of scoring data in the objective function is:
error=weight(rui-predict)
Wherein r ui represents the security grade score of the ith row and the ith column of the scoring matrix, predict represents the predicted security grade score, and the predicted security grade score is obtained by calculating the inner product of the ith row and the ith column of the scoring matrix. In the case of considering the scoring weight of the scoring data, the weight error value of each piece of scoring data is:
WeightError=weight*error
wherein weight is the scoring weight of the scoring data. The machine learning model will minimize the weight error values for all scoring data.
And after the grading data of all the users are input into the machine learning model for parameter training, a food similarity matrix W is obtained. The predicted safety rating score value of the unscored food i of the user u is calculated as follows:
PredictRateui＝RuWi
Wherein R u represents the vector of the ith row in the scoring matrix, namely the scoring data of the user u, and W i represents the vector of the ith column in the food similarity matrix, namely the similarity value of the food i and other foods. After calculating the predicted safety grade grading values of all unscored foods by the user, sorting the predicted safety grade grading values, and selecting a plurality of foods with highest grading values for recommending to the user. And finally, calculating and generating a food recommendation list of the user through the scoring data of the user and the food similarity matrix, and recommending the food to the user.
In summary, after the scheme is adopted, the method breaks through the problems of low recommending effect based on the neighborhood method and low recommending efficiency based on the model method, and calculates the scoring weight under the condition of considering the time sequence of the scoring data of the user, so that the method can effectively generate the recommendation, and the recommending effect is improved.
The above embodiments are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention, so variations in shape and principles of the present invention should be covered.
Claims (3)
1. A recommendation method based on a user's food safety rating score and food similarity, comprising the steps of:
1) The method comprises the steps of obtaining grading data of a user on food, wherein each piece of grading data comprises a user ID, food ID, a security grade grading value and grading time;
2) For scoring data of the same user ID, calculating the scoring weight of each piece of scoring data according to the sequence before and after the scoring time and the security grade scoring value;
3) Inputting the user ID of the grading data, the food ID, the safety grade grading value and the grading weight into a machine learning model for parameter training, obtaining a food similarity matrix after the parameter training is completed, and finally calculating and generating a food recommendation list of the user through the grading data of the user and the food similarity matrix to realize food recommendation to the user, wherein the food recommendation list comprises the following specific steps:
The method comprises the steps of converting scoring data of all users into a scoring matrix, wherein each row of the scoring matrix represents one user, each column of the scoring matrix represents one food, each value of the matrix represents the security grade scoring value of the row of the food by the user, inputting the scoring matrix into a machine learning model for parameter training, and the matrix form objective function of the machine learning model is as follows:
Wherein R is a scoring matrix of m rows and n columns, m represents the number of users, n represents the number of foods, each row of the scoring matrix represents a user, each column represents a food, W is a training-obtained n-row and n-column food similarity matrix, n represents the number of foods, and each row or each column represents a food, each value in the food similarity matrix represents the similarity of the row of foods and the column of foods, the value is greater than or equal to zero, and meanwhile, the value on the diagonal of the food similarity matrix is zero, the L1 norm of the matrix is equal to the L 1, the L F is the Frobenius norm of the matrix, and the constants beta and lambda are regularization parameters;
Since each column in the food similarity matrix W is independent, each column of the food similarity matrix can be trained in parallel, and the matrix-form objective function can be decomposed into:
Wherein R j is the j-th column of the scoring matrix R, W j is the j-th column of the food similarity matrix W, I I.I. 2 is the L2 norm of the vector, I I.I. 1 is the L1 norm of the vector, constants beta and lambda are regularization parameters, in order to reduce the training time of model parameters, the safety grade scoring value of each piece of scoring data is multiplied by the corresponding scoring weight, a neighborhood-based collaborative filtering method is used for calculating the column with high similarity between R j and the scoring matrix, and only the column with high similarity between R j is selected to fit an objective function in the minimization of the objective function;
the error value formula for calculating each piece of scoring data in the objective function is:
error=weight(rui-predict)
Wherein r ui represents the security grade score value of the ith row and the ith column of the scoring matrix, predict represents the predicted security grade score value, the predicted security grade score value is obtained by calculating the inner product of the ith row of the scoring matrix and the ith column of the food similarity matrix, and the weight error value of each piece of scoring data is as follows under the condition of considering the scoring weight of the scoring data:
WeightError=weight*error
The machine learning model minimizes the weight error value of all the scoring data;
After the scoring data of all users are input into the machine learning model for parameter training, a food similarity matrix W is obtained, and the predicted safety grade scoring value of the user u on the unscored food i is calculated as follows:
PredictRateui＝RuWi
the method comprises the steps of calculating predicted safety grade grading values of all unscored foods by a user, sorting the predicted safety grade grading values, selecting a plurality of foods with highest grading values for recommending to the user, and completing the recommendation process, wherein R u represents a vector of a u th row in a grading matrix, namely grading data of the user u, W i represents a vector of an i th column in a food similarity matrix, namely similarity values of the food i and other foods.
2. The recommendation method based on food safety grade grading values and food similarity of users according to claim 1, wherein in the step 1), grading data of the users on food refer to historical grading data of the users for inputting a machine learning model for training, wherein the grading data comprise user IDs, food IDs, safety grade grading values and grading time, the user IDs are unique identifiers of the users, different user IDs represent different users, the food IDs are unique identifiers of the food, different food IDs represent different food, the safety grade grading values refer to the safety grade grading values of the food represented by the user IDs, and the grading time refers to the grading time of the user on the food safety grade.
3. The recommendation method based on the food safety rating score and the food similarity of claim 1, wherein in the step 2), the rating data of different users are grouped according to the user ID, the rating data of the same user ID are arranged in sequence according to the rating time, and the rating weight value of each rating data after the sequencing is processed according to three conditions that ① if the safety rating score of the previous rating data is the highest rating value, the rating weight of the current rating data is 1.2, and ② if the safety rating score of the previous rating data is the lowest rating value, the rating weight of the current rating data is 0.8, and the rating weight of the first rating data and other data of ③ is 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111003703.6A CN113888252B (en) | 2021-08-30 | 2021-08-30 | Recommendation method based on user's food safety rating and food similarity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111003703.6A CN113888252B (en) | 2021-08-30 | 2021-08-30 | Recommendation method based on user's food safety rating and food similarity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113888252A CN113888252A (en) | 2022-01-04 |
| CN113888252B true CN113888252B (en) | 2025-01-10 |
Family
ID=79011686
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111003703.6A Active CN113888252B (en) | 2021-08-30 | 2021-08-30 | Recommendation method based on user's food safety rating and food similarity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113888252B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106682121A (en) * | 2016-12-09 | 2017-05-17 | 广东工业大学 | Time utility recommendation method based on interest change of user |
| CN111414555A (en) * | 2020-01-06 | 2020-07-14 | 浙江工业大学 | A Personalized Recommendation Method Based on Collaborative Filtering |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7225182B2 (en) * | 1999-05-28 | 2007-05-29 | Overture Services, Inc. | Recommending search terms using collaborative filtering and web spidering |
| EP2627096A1 (en) * | 2012-02-09 | 2013-08-14 | Thomson Licensing | Recommendation method using similarity metrics |
| JP5945206B2 (en) * | 2012-10-02 | 2016-07-05 | 日本電信電話株式会社 | Product recommendation device, method and program |
| CN104281956B (en) * | 2014-10-27 | 2018-09-07 | 南京信息工程大学 | The dynamic recommendation method for adapting to user interest variation based on temporal information |
| CN105354330A (en) * | 2015-11-27 | 2016-02-24 | 南京邮电大学 | Sparse data preprocessing based collaborative filtering recommendation method |
| CN109543109B (en) * | 2018-11-27 | 2021-06-22 | 山东建筑大学 | A Recommendation Algorithm Integrating Time Window Technology and Rating Prediction Model |
-
2021
- 2021-08-30 CN CN202111003703.6A patent/CN113888252B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106682121A (en) * | 2016-12-09 | 2017-05-17 | 广东工业大学 | Time utility recommendation method based on interest change of user |
| CN111414555A (en) * | 2020-01-06 | 2020-07-14 | 浙江工业大学 | A Personalized Recommendation Method Based on Collaborative Filtering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113888252A (en) | 2022-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110110181B (en) | A clothing matching recommendation method based on user style and scene preference | |
| Geng et al. | Early warning modeling and analysis based on analytic hierarchy process integrated extreme learning machine (AHP-ELM): Application to food safety | |
| CN110362738B (en) | Deep learning-based individual recommendation method combining trust and influence | |
| CN109241366B (en) | A hybrid recommender system based on multi-task deep learning and its method | |
| CN109785062A (en) | A kind of hybrid neural networks recommender system based on collaborative filtering model | |
| CN106779467A (en) | Enterprises ' industry categorizing system based on automatic information screening | |
| CN111260201B (en) | Variable importance analysis method based on layered random forest | |
| CN105787100A (en) | User session recommendation method based on deep neural network | |
| CN112818256B (en) | A recommendation method based on neural collaborative filtering | |
| CN103886486A (en) | Electronic commerce recommending method based on support vector machine (SVM) | |
| CN106339718A (en) | Classification method based on neural network and classification device thereof | |
| CN104268572B (en) | Feature extraction and feature selection approach towards backstage multi-source data | |
| CN108960304A (en) | A kind of deep learning detection method of network trading fraud | |
| CN117611273B (en) | Cross-domain recommendation method based on source domain data enhancement and multi-interest refinement transfer | |
| CN113763031A (en) | Commodity recommendation method and device, electronic equipment and storage medium | |
| CN113159892A (en) | Commodity recommendation method based on multi-mode commodity feature fusion | |
| CN111966888A (en) | External data fused interpretable recommendation method and system based on aspect categories | |
| Miao et al. | A recommendation system based on text mining | |
| Farahani et al. | Car sales forecasting using artificial neural networks and analytical hierarchy process | |
| Lei et al. | Composing recipes based on nutrients in food in a machine learning context | |
| CN111583363A (en) | Visual automatic generation method and system for image-text news | |
| CN113888252B (en) | Recommendation method based on user's food safety rating and food similarity | |
| CN104572915A (en) | User event relevance calculation method based on content environment enhancement | |
| CN113487377A (en) | Individualized real-time recommendation method based on GRU network | |
| CN114997959A (en) | Electronic intelligent product marketing recommendation method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |