CN118172138A - Intelligent recommendation system for electronic commerce - Google Patents

Intelligent recommendation system for electronic commerce Download PDF

Info

Publication number
CN118172138A
CN118172138A CN202410280498.5A CN202410280498A CN118172138A CN 118172138 A CN118172138 A CN 118172138A CN 202410280498 A CN202410280498 A CN 202410280498A CN 118172138 A CN118172138 A CN 118172138A
Authority
CN
China
Prior art keywords
user
commodity
recommendation
module
data
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
Application number
CN202410280498.5A
Other languages
Chinese (zh)
Inventor
张禹
潘银芳
徐兴华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Industry And Trade Vocational And Technical College Zhejiang Industry And Trade Technician College
Original Assignee
Zhejiang Industry And Trade Vocational And Technical College Zhejiang Industry And Trade Technician College
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang Industry And Trade Vocational And Technical College Zhejiang Industry And Trade Technician College filed Critical Zhejiang Industry And Trade Vocational And Technical College Zhejiang Industry And Trade Technician College
Priority to CN202410280498.5A priority Critical patent/CN118172138A/en
Publication of CN118172138A publication Critical patent/CN118172138A/en
Pending legal-status Critical Current

Links

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The application discloses an intelligent recommendation system for electronic commerce, which relates to the technical field of electronic commerce and comprises the following components: the system comprises a user data acquisition module, a user management module, a commodity management module and a purchase recommendation module; the user data acquisition module is used for acquiring behavior data, historical shopping records and user basic information of a user and preprocessing the data information; the user management module is used for verifying and managing the user basic information and modeling and managing the user files and the user commodity orders; the commodity management module is used for carrying out class modeling on commodities according to commodity classes, carrying out commodity retrieval according to the search key words and taking part of characters as query conditions, carrying out update management on the commodities and displaying the commodities; the purchase recommendation module is used for recommending personalized commodities to the user by adopting the hybrid intelligent recommendation model, the cold start problem of the system is effectively relieved, and the recommendation prediction accuracy is higher by adopting the hybrid recommendation algorithm.

Description

Intelligent recommendation system for electronic commerce
Technical Field
The application relates to the technical field of electronic commerce, in particular to an intelligent recommendation system for electronic commerce.
Background
With the development of internet economy, people increasingly select online shopping. The development of electronic commerce enables people to live conveniently, quickly and in various colors, and most of the income of electronic commerce websites benefits from the recommendation engine deployed by the online system, and the recommendation method can help the user group to accurately find target commodities in a short time, so that the user experience is improved. Most of the existing recommendation systems are built based on statistical knowledge and machine learning methods, and the histories of users are processed to obtain user figures, and then the features are input into a recommendation system model to be operated to generate a result. However, the method cannot dynamically apply feedback information of the user, so that the recommendation effect is lagged and the efficiency is low.
Disclosure of Invention
In view of the above, the application provides an intelligent recommendation system for electronic commerce, which can improve recommendation accuracy and recommendation efficiency and avoid the technical problems of insufficient new user behavior data and inaccurate extraction of user and commodity characteristics during cold start.
In order to achieve the above purpose, the present invention provides the following technical solutions:
The intelligent recommendation system for electronic commerce comprises: the system comprises a user data acquisition module, a user management module, a commodity management module and a purchase recommendation module;
the user data acquisition module is used for acquiring behavior data, historical shopping records and user basic information of a user and preprocessing the acquired data information;
The user management module is connected with the user data acquisition module and is used for verifying and managing the user basic information and modeling and managing the user files and the user commodity orders;
The commodity management module is used for carrying out class modeling on commodities according to commodity classes, carrying out commodity retrieval according to the search keyword and taking part of characters as query conditions, carrying out update management on the commodities, and displaying the commodities according to the purchase recommendation result;
the purchase recommendation module is connected with the user management module and the commodity modeling module, and is used for conducting personalized commodity recommendation on the user by adopting a hybrid intelligent recommendation model.
Further, the user data acquisition module comprises a data acquisition module and a preprocessing module;
The data acquisition module acquires basic information and behavior data of a user through user input data, webpage comments and log information, and manages the basic information and the behavior data of the user to form a user file;
The preprocessing module is connected with the data acquisition module and is used for carrying out association processing on user behavior data through four groups of user name, commodity name, operation and time and storing the preprocessed data into a database.
Further, the user management module comprises a registration login module, a user information changing module and an order management module; the registration login module is used for verifying and managing the user identity, matching corresponding operation authorities and network resources for different users, and forming a user file; the user information changing module is used for acquiring basic information and application requests of users, carrying out identity verification and information changing according to the application requests and changing limiting conditions of the users, and storing the user information into the user database; the order management module is used for storing and deleting the transaction order data of the user.
Further, the purchase recommendation module comprises a user classification analysis module and a recommendation calculation module;
The user classification analysis module is used for judging the user category according to the user login information, classifying the different user categories according to the judgment result, matching the corresponding recommendation model for the user based on the label, and obtaining a commodity recommendation result based on the corresponding recommendation model;
the recommendation calculation module is connected with the user classification analysis module and is used for conducting personalized recommendation on the user according to an intelligent recommendation algorithm and user classification information.
Still further, the personalized commodity recommendation to the user by adopting the hybrid intelligent recommendation model comprises:
acquiring user classification information, and determining user categories, wherein the user categories comprise new users and non-new users;
And carrying out commodity recommendation on different types of users by adopting a mixed method, wherein the mixed method comprises a commodity recommendation algorithm based on a time factor and a commodity recommendation algorithm based on user interestingness, wherein the commodity recommendation algorithm based on the time factor is adopted to carry out commodity recommendation when the grading filling value of the system is larger than a preset threshold value, and the commodity recommendation algorithm based on the user interestingness is adopted to carry out commodity recommendation when the grading filling value of the system is smaller than the preset threshold value.
Still further, the commodity recommendation based on the time factor commodity recommendation algorithm comprises:
acquiring user data, commodity data, historical shopping records and scoring data, and calculating the number of users m, the number of commodities n and a user scoring matrix R m×n according to the user data, the commodity data and the scoring data;
Calculating the paranoid of a user e according to the number m of users, the time factor and the user scoring matrix, calculating the paranoid of a commodity I according to the number n of commodities, the time factor and the user scoring matrix, inputting the calculated user paranoid, commodity paranoid, user scoring and user behavior data into a prediction model of the commodity by the user to obtain a user characteristic U and a commodity characteristic I, and training the prediction model by adopting a gradient descent method;
calculating the predicted score of the commodity by a user according to the commodity characteristic I by adopting a cosine similarity calculation method, and carrying out reverse sequencing on the predicted score to output a primary commodity recommendation list;
And acquiring the latest W times of historical shopping records of the user, calculating a new similarity score based on the historical shopping records and the primary commodity recommendation list, and outputting a final splash recommendation list according to the score sorting result.
Still further, the commodity recommendation algorithm based on the user interest level comprises:
Acquiring browsing, collecting, sharing and purchasing behavior data of a user and commodity comment data according to the user behavior data, calculating the interest weight of the user on the commodity according to the browsing, collecting, sharing and purchasing behavior data of the user to obtain a user interest weight grade A1, and carrying out iterative updating on the user interest weight grade A1 according to time attenuation and behaviors;
Analyzing the emotion level A2 of the user on the commodity by adopting a text emotion analysis method according to commodity comment data of the user, and analyzing the interest degree B=θA1+ (1-theta) A2 of the user on the commodity according to the emotion level A2 of the commodity;
Filling the sparse scoring matrix by adopting interestingness, decomposing the scoring matrix to obtain a user characteristic matrix and a commodity characteristic matrix, calculating to obtain a scoring prediction matrix by adopting a collaborative filtering method based on the user characteristic matrix and the commodity characteristic matrix, and sequencing scoring results in a descending order to generate a personalized commodity recommendation list.
Further, the analyzing the emotion level A2 of the user to the commodity by adopting the text emotion analysis method comprises the following steps:
preprocessing the obtained comment text information through a text model to obtain a Word sequence, and obtaining Word vectors based on the Word sequence by adopting a Word embedding model Word2 vec;
training word vectors by adopting an emotion text analysis method based on an LSTM network to obtain a prediction grade A2 based on emotion analysis.
Further, the commodity management module comprises a commodity inquiry module, a management module and a commodity display module; the commodity inquiry module is used for searching commodities of corresponding classes according to inquiry information input by a user; the management module is used for maintaining the commodities by an administrator, and carrying out operations of loading the commodities, updating the quantity of the commodities, deleting the commodities and modifying the commodity information;
The commodity display module is used for modeling commodities according to commodity categories to form a commodity list, and displaying and recommending the commodities according to the commodity list and the recommendation result.
From the above technical solution, the advantages of the present invention are:
1. The method effectively relieves the cold start problem of the system, fully mines and analyzes the user data, adopts the mixed recommendation algorithm in consideration of time influence factors to ensure that the recommendation prediction accuracy is higher, and timely recommendation results are obtained, thereby greatly improving the shopping experience of the user.
2. By adopting the mode of filling the scoring matrix by the user interestingness, the problem of low feature accuracy of user features, commodity features and the like caused by sparse scoring matrix in the traditional recommendation algorithm is avoided.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application.
FIG. 1 is a schematic diagram of the composition structure of an intelligent recommendation system for electronic commerce according to the present application.
Fig. 2 is a schematic diagram of another composition structure of an intelligent recommendation system for electronic commerce according to the present embodiment.
Fig. 3 is a schematic diagram of specific steps of personalized commodity recommendation for a user by using a hybrid intelligent recommendation model in this embodiment.
Fig. 4 is a schematic diagram of specific steps of commodity recommendation performed by the commodity recommendation algorithm based on the time factor in the present embodiment.
Fig. 5 is a schematic diagram of specific steps of commodity recommendation performed by the commodity recommendation algorithm based on the user interest level in the present embodiment.
Detailed Description
The present application will be described in further detail with reference to the following embodiments and the accompanying drawings, in order to make the objects, technical solutions and advantages of the present application more apparent. The exemplary embodiments of the present application and the descriptions thereof are used herein to explain the present application, but are not intended to limit the application.
A variety of recommendation systems are used for many large online shopping sites, such as naughty net, jindong mall, amazon, etc. For each shopping mall and website, a good recommendation system can improve shopping experience of users, attract more user resources and improve enterprise competitiveness. Most of the current recommendation systems are built based on statistical knowledge and machine learning methods, and the recommendation methods lack dynamic feedback recommendation. Referring to fig. 1 to 5, the present embodiment provides an intelligent recommendation system for electronic commerce, which can solve the problem of cold start of the system, and adopts a hybrid recommendation algorithm in consideration of time influence factors to make recommendation prediction accuracy higher, because closer purchased commodities reflect the recent preference of users, and farther purchased commodities have smaller influence factors and can manage different users, the intelligent recommendation system for electronic commerce comprises: the system comprises a user data acquisition module, a user management module, a commodity management module and a purchase recommendation module. The user data acquisition module is used for acquiring behavior data, historical shopping records and user basic information of a user and preprocessing the acquired data information.
Specifically, the user data acquisition module comprises a data acquisition module and a preprocessing module; the data acquisition module acquires basic information and behavior data of a user through user input data, webpage comments and log information, and manages the basic information and the behavior data of the user to form a user file; the preprocessing module is connected with the data acquisition module and is used for carrying out association processing on user behavior data through four groups of user name, commodity name, operation and time and storing the preprocessed data into a database.
In this embodiment, the personal information of the user mainly includes information such as a user name, a user gender, a birthday, a mailbox, an address where the user is located, and the like. The behavior data of the user comprise user browsing data, clicking data, operation data, storage data and the like, and the browsing and clicking behaviors of the user can be collected through the log embedded point.
The user management module is connected with the user data acquisition module and is used for verifying and managing the user basic information and modeling and managing the user files and the user commodity orders.
The user management module comprises a registration login module, a user information changing module and an order management module; the registration login module is used for verifying and managing the user identity, matching corresponding operation authorities and network resources for different users, and forming a user file; the user information changing module is used for acquiring basic information and application requests of users, carrying out identity verification and information changing according to the application requests and changing limiting conditions of the users, and storing the user information into the user database; the order management module is used for storing and deleting the transaction order data of the user.
In this embodiment, when a new user enters the system, the user identity needs to be registered, when the user logs in, the user identity is verified according to the registration information, the user login mode includes password login and the like, when the user logs in the system, the user name and the password are input, the corresponding user is queried through the user name and the password, if the query is empty, the user name or the password is wrong, if the query result is not empty, the user type is judged, if the user is a common user, the user jumps to a shopping page at the front end of the system, otherwise, the user is an administrator, and jumps to a commodity management system. Besides the common users, the system also comprises an administrator for maintaining commodities, and operations such as loading, quantity updating, deleting, commodity information modification and the like are performed on the commodities, so that operation authorities and network resource allocation and information storage according to the identification are required to be performed on different users.
The commodity management module is used for carrying out class modeling on commodities according to commodity classes, carrying out commodity retrieval according to the search keyword and taking part of characters as query conditions, carrying out updating management on the commodities, and displaying the commodities according to the purchase recommendation result.
The commodity management module comprises a commodity inquiry module, a management module and a commodity display module; the commodity inquiry module is used for searching commodities of corresponding classes according to inquiry information input by a user; the management module is used for maintaining the commodities by an administrator, and carrying out operations of loading the commodities, updating the quantity of the commodities, deleting the commodities and modifying the commodity information; the commodity display module is used for modeling commodities according to commodity categories to form a commodity list, and displaying and recommending the commodities according to the commodity list and the recommendation result.
In this embodiment, the commodity table is used to store corresponding data of commodities, including commodity names, directory names, picture addresses, labels, and the like.
The purchase recommendation module is connected with the user management module and the commodity modeling module, and is used for conducting personalized commodity recommendation on the user by adopting a hybrid intelligent recommendation model.
The purchase recommendation module comprises a user classification analysis module and a recommendation calculation module; the user classification analysis module is used for judging the user category according to the user login information, classifying the different user categories according to the judgment result, matching the corresponding recommendation model for the user based on the label, and obtaining a commodity recommendation result based on the corresponding recommendation model; the recommendation calculation module is connected with the user classification analysis module and is used for conducting personalized recommendation on the user according to an intelligent recommendation algorithm and user classification information.
Specifically, the personalized commodity recommendation for the user by adopting the hybrid intelligent recommendation model comprises the following steps:
step S1: acquiring user classification information, and determining user categories, wherein the user categories comprise new users and non-new users;
Step S2: and carrying out commodity recommendation on different types of users by adopting a mixed method, wherein the mixed method comprises a commodity recommendation algorithm based on a time factor and a commodity recommendation algorithm based on user interestingness, wherein the commodity recommendation algorithm based on the time factor is adopted to carry out commodity recommendation when the grading filling value of the system is larger than a preset threshold value, and the commodity recommendation algorithm based on the user interestingness is adopted to carry out commodity recommendation when the grading filling value of the system is smaller than the preset threshold value.
In this embodiment, different recommendation algorithms are adopted to recommend the goods, so that switching calculation is required for different states, the key of algorithm switching is the scoring filling threshold of the system, the threshold can be configured by a manager, when the scoring filling value of the system is greater than a preset threshold, the scoring data in the system is sufficient, the commodity recommendation algorithm based on a time factor is adopted to recommend the goods, the time factor is integrated into the recommendation algorithm to predict 'real-time', and the acquired user characteristics and commodity characteristics are more accurate. When the score filling value of the system is smaller than a preset threshold value, the score data in the system are relatively less, so that commodity recommendation is required to be performed by adopting a commodity recommendation algorithm based on the user interestingness.
More specifically, the commodity recommendation based on the time factor commodity recommendation algorithm comprises the following steps:
Step A: acquiring user data, commodity data, historical shopping records and scoring data, and calculating the number of users m, the number of commodities n and a user scoring matrix R m×m according to the user data, the commodity data and the scoring data;
And (B) step (B): calculating the paranoid of a user e according to the number m of users, the time factor and the user scoring matrix, calculating the paranoid of a commodity I according to the number n of commodities, the time factor and the user scoring matrix, inputting the calculated user paranoid, commodity paranoid, user scoring and user behavior data into a prediction model of the commodity by the user to obtain a user characteristic U and a commodity characteristic I, and training the prediction model by adopting a gradient descent method;
step C: calculating the predicted score of the commodity by a user according to the commodity characteristic I by adopting a cosine similarity calculation method, and carrying out reverse sequencing on the predicted score to output a primary commodity recommendation list;
Step D: and acquiring the latest W times of historical shopping records of the user, calculating a new similarity score based on the historical shopping records and the primary commodity recommendation list, and outputting a final splash recommendation list according to the score sorting result.
In this embodiment, the user employs a conventional predictive scoring model that incorporates a temporal penalty factor to the predictive model of the good.
The commodity recommendation based on the commodity recommendation algorithm of the user interest degree comprises the following steps:
Step A: acquiring browsing, collecting, sharing and purchasing behavior data of a user and commodity comment data according to the user behavior data, calculating the interest weight of the user on the commodity according to the browsing, collecting, sharing and purchasing behavior data of the user to obtain a user interest weight grade A1, and carrying out iterative updating on the user interest weight grade A1 according to time attenuation and behaviors;
and (B) step (B): analyzing emotion level A2 of the user on the commodity by adopting a text emotion analysis method according to commodity comment data of the user, and analyzing interest degree B=θA1+ (1- θ) A2 of the user on the commodity according to emotion level A2 of the commodity, wherein θ is a weight factor;
Step C: filling the sparse scoring matrix by adopting interestingness, decomposing the scoring matrix to obtain a user characteristic matrix and a commodity characteristic matrix, calculating to obtain a scoring prediction matrix by adopting a collaborative filtering method based on the user characteristic matrix and the commodity characteristic matrix, and sequencing scoring results in a descending order to generate a personalized commodity recommendation list.
In this embodiment, the conventional recommendation algorithm is mainly implemented by collaborative filtering, only a scoring matrix of the commodity by the user is utilized, and when the scoring matrix is sparse, the obtained user feature vector and commodity feature vector are often not high in accuracy, the behavior of the user on the commodity not only comprises scoring, but also has higher mining value of other behavior data, including browsing, shopping cart adding, collection and the like of the commodity, and the preference degree of the user on the commodity can be reflected.
More specifically, the analyzing the emotion level A2 of the user to the commodity by adopting the text emotion analysis method comprises the following steps:
Step B1: preprocessing the obtained comment text information through a text model to obtain a Word sequence, and obtaining Word vectors based on the Word sequence by adopting a Word embedding model Word2 vec;
Step B2: training word vectors by adopting an emotion text analysis method based on an LSTM network to obtain a prediction grade A2 based on emotion analysis.
The above description is only of the preferred embodiments of the present application and is not intended to limit the present application, and various modifications and variations can be made to the embodiments of the present application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims (9)

1. An intelligent recommendation system for electronic commerce, comprising: the system comprises a user data acquisition module, a user management module, a commodity management module and a purchase recommendation module;
the user data acquisition module is used for acquiring behavior data, historical shopping records and user basic information of a user and preprocessing the acquired data information;
The user management module is connected with the user data acquisition module and is used for verifying and managing the user basic information and modeling and managing the user files and the user commodity orders;
The commodity management module is used for carrying out class modeling on commodities according to commodity classes, carrying out commodity retrieval according to the search keyword and taking part of characters as query conditions, carrying out update management on the commodities, and displaying the commodities according to the purchase recommendation result;
the purchase recommendation module is connected with the user management module and the commodity modeling module, and is used for conducting personalized commodity recommendation on the user by adopting a hybrid intelligent recommendation model.
2. The intelligent recommendation system for electronic commerce according to claim 1, wherein the user data acquisition module comprises a data acquisition module and a preprocessing module;
The data acquisition module acquires basic information and behavior data of a user through user input data, webpage comments and log information, and manages the basic information and the behavior data of the user to form a user file;
The preprocessing module is connected with the data acquisition module and is used for carrying out association processing on user behavior data through four groups of user name, commodity name, operation and time and storing the preprocessed data into a database.
3. The intelligent recommendation system for electronic commerce according to claim 2, wherein the user management module comprises a registration login module, a user information modification module, and an order management module; the registration login module is used for verifying and managing the user identity, matching corresponding operation authorities and network resources for different users, and forming a user file; the user information changing module is used for acquiring basic information and application requests of users, carrying out identity verification and information changing according to the application requests and changing limiting conditions of the users, and storing the user information into the user database; the order management module is used for storing and deleting the transaction order data of the user.
4. The intelligent recommendation system for electronic commerce according to claim 1, wherein the purchase recommendation module comprises a user classification analysis module and a recommendation calculation module;
The user classification analysis module is used for judging the user category according to the user login information, classifying the different user categories according to the judgment result, matching the corresponding recommendation model for the user based on the label, and obtaining a commodity recommendation result based on the corresponding recommendation model;
the recommendation calculation module is connected with the user classification analysis module and is used for conducting personalized recommendation on the user according to an intelligent recommendation algorithm and user classification information.
5. The intelligent recommendation system for electronic commerce according to claim 4, wherein said employing a hybrid intelligent recommendation model to conduct personalized commodity recommendation for a user comprises:
acquiring user classification information, and determining user categories, wherein the user categories comprise new users and non-new users;
And carrying out commodity recommendation on different types of users by adopting a mixed method, wherein the mixed method comprises a commodity recommendation algorithm based on a time factor and a commodity recommendation algorithm based on user interestingness, wherein the commodity recommendation algorithm based on the time factor is adopted to carry out commodity recommendation when the grading filling value of the system is larger than a preset threshold value, and the commodity recommendation algorithm based on the user interestingness is adopted to carry out commodity recommendation when the grading filling value of the system is smaller than the preset threshold value.
6. The intelligent recommendation system for electronic commerce according to claim 5, wherein said time-factor-based commodity recommendation algorithm makes commodity recommendations comprising:
acquiring user data, commodity data, historical shopping records and scoring data, and calculating the number of users m, the number of commodities n and a user scoring matrix R m×n according to the user data, the commodity data and the scoring data;
Calculating the paranoid of a user e according to the number m of users, the time factor and the user scoring matrix, calculating the paranoid of a commodity I according to the number n of commodities, the time factor and the user scoring matrix, inputting the calculated user paranoid, commodity paranoid, user scoring and user behavior data into a prediction model of the commodity by the user to obtain a user characteristic U and a commodity characteristic I, and training the prediction model by adopting a gradient descent method;
calculating the predicted score of the commodity by a user according to the commodity characteristic I by adopting a cosine similarity calculation method, and carrying out reverse sequencing on the predicted score to output a primary commodity recommendation list;
And acquiring the latest W times of historical shopping records of the user, calculating a new similarity score based on the historical shopping records and the primary commodity recommendation list, and outputting a final splash recommendation list according to the score sorting result.
7. The intelligent recommendation system for electronic commerce according to claim 5, wherein the commodity recommendation algorithm based on the user interest level comprises:
Acquiring browsing, collecting, sharing and purchasing behavior data of a user and commodity comment data according to the user behavior data, calculating the interest weight of the user on the commodity according to the browsing, collecting, sharing and purchasing behavior data of the user to obtain a user interest weight grade A1, and carrying out iterative updating on the user interest weight grade A1 according to time attenuation and behaviors;
Analyzing the emotion level A2 of the user on the commodity by adopting a text emotion analysis method according to commodity comment data of the user, and analyzing the interest degree B=θA1+ (1-theta) A2 of the user on the commodity according to the emotion level A2 of the commodity;
Filling the sparse scoring matrix by adopting interestingness, decomposing the scoring matrix to obtain a user characteristic matrix and a commodity characteristic matrix, calculating to obtain a scoring prediction matrix by adopting a collaborative filtering method based on the user characteristic matrix and the commodity characteristic matrix, and sequencing scoring results in a descending order to generate a personalized commodity recommendation list.
8. The intelligent recommendation system for electronic commerce according to claim 7, wherein said analyzing emotion level A2 of a user to a commodity using a text emotion analysis method comprises:
preprocessing the obtained comment text information through a text model to obtain a Word sequence, and obtaining Word vectors based on the Word sequence by adopting a Word embedding model Word2 vec;
training word vectors by adopting an emotion text analysis method based on an LSTM network to obtain a prediction grade A2 based on emotion analysis.
9. The intelligent recommendation system for electronic commerce according to claim 1, wherein the commodity management module comprises a commodity inquiry module, a management module and a commodity display module; the commodity inquiry module is used for searching commodities of corresponding classes according to inquiry information input by a user; the management module is used for maintaining the commodities by an administrator, and carrying out operations of loading the commodities, updating the quantity of the commodities, deleting the commodities and modifying the commodity information; the commodity display module is used for modeling commodities according to commodity categories to form a commodity list, and displaying and recommending the commodities according to the commodity list and the recommendation result.
CN202410280498.5A 2024-03-12 2024-03-12 Intelligent recommendation system for electronic commerce Pending CN118172138A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410280498.5A CN118172138A (en) 2024-03-12 2024-03-12 Intelligent recommendation system for electronic commerce

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410280498.5A CN118172138A (en) 2024-03-12 2024-03-12 Intelligent recommendation system for electronic commerce

Publications (1)

Publication Number Publication Date
CN118172138A true CN118172138A (en) 2024-06-11

Family

ID=91357630

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410280498.5A Pending CN118172138A (en) 2024-03-12 2024-03-12 Intelligent recommendation system for electronic commerce

Country Status (1)

Country Link
CN (1) CN118172138A (en)

Similar Documents

Publication Publication Date Title
CN107391687B (en) Local log website-oriented hybrid recommendation system
CN110704572B (en) Suspected illegal fundraising risk early warning method, device, equipment and storage medium
AU2022201654A1 (en) System and engine for seeded clustering of news events
CN102483745B (en) Co-selected image classification
CN111708949B (en) Medical resource recommendation method and device, electronic equipment and storage medium
US7580926B2 (en) Method and apparatus for representing text using search engine, document collection, and hierarchal taxonomy
CN108363821A (en) A kind of information-pushing method, device, terminal device and storage medium
CN107862553A (en) Advertisement real-time recommendation method, device, terminal device and storage medium
US20090327249A1 (en) Intellegent Data Search Engine
CN111144723A (en) Method and system for recommending people's job matching and storage medium
CN106844407B (en) Tag network generation method and system based on data set correlation
CN112632405B (en) Recommendation method, recommendation device, recommendation equipment and storage medium
CN111125343A (en) Text analysis method and device suitable for human-sentry matching recommendation system
CN101520784A (en) Information issuing system and information issuing method
CN115002200B (en) Message pushing method, device, equipment and storage medium based on user portrait
CN103577489A (en) Method and device of searching web browsing history
CN111105209A (en) Job resume matching method and device suitable for post matching recommendation system
CN114238573B (en) Text countercheck sample-based information pushing method and device
CN103577490A (en) Method and device of showing web browsing history
CN115659008B (en) Information pushing system, method, electronic equipment and medium for big data information feedback
CN110637316A (en) System and method for intelligent prospective object recognition using online resources and neural network processing to classify tissue based on published material
CN114254201A (en) Recommendation method for science and technology project review experts
CA2956627A1 (en) System and engine for seeded clustering of news events
CN113420018A (en) User behavior data analysis method, device, equipment and storage medium
CN111310032A (en) Resource recommendation method and device, computer equipment and readable storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination