Search

CN-121981794-A - Commodity recommendation method and system based on consumer behaviors

CN121981794ACN 121981794 ACN121981794 ACN 121981794ACN-121981794-A

Abstract

The application provides a commodity recommendation method and a commodity recommendation system based on consumer behaviors, which are characterized by firstly extracting explicit behavior features and implicit preference features of a consumer in different context scenes, determining node preference indexes of the consumer behaviors in different context scenes, generating a short-term intention chain and a long-term preference chain of the consumer according to the node preference indexes, further constructing element feature maps reflecting consumer behavior preferences according to mapping association relations between the short-term intention chain and the long-term preference chain, extracting preference weights of preference nodes in the element feature maps in different context scenes, detecting the matching degree of the element feature maps and current consumption behavior features, and updating the element feature maps according to the preference weights corresponding to the preference nodes and the current consumption behavior features when the matching degree is lower than a preset threshold value, so as to generate a commodity recommendation list. By adopting the scheme of the application, personalized commodity recommendation based on different context scenes can be realized in the Internet E-commerce platform.

Inventors

  • XIE JINGJING

Assignees

  • 海南外国语职业学院

Dates

Publication Date
20260505
Application Date
20251224

Claims (10)

  1. 1. A commodity recommendation method based on consumer behavior, comprising the steps of: Collecting consumption behavior data of a consumer in an Internet e-commerce platform, and further extracting explicit behavior characteristics and implicit preference characteristics of the consumer in different context scenes; Performing behavior node association on the explicit behavior features and the implicit preference features under different context scenes to obtain node preference indexes of the consumption behaviors of the consumer under different context scenes; Generating a short-term intention chain and a long-term preference chain of a consumer according to all node preference indexes and the time sequence corresponding to each context scene, and further constructing an element feature map reflecting consumer consumption behavior preference according to the mapping incidence relation between the short-term intention chain and the long-term preference chain; extracting preference weights of each preference node in the element feature map under different context scenes; Detecting the matching degree of the element feature map and the current consumption behavior feature, and when the matching degree is lower than a preset threshold value, updating the element feature map in a preference mode based on the preference weight corresponding to each preference node and the current consumption behavior feature, and further generating a commodity recommendation list based on the updated element feature map.
  2. 2. The method of claim 1, wherein extracting explicit behavior features and implicit preference features of a consumer in different context scenarios specifically comprises: acquiring consumption behavior data of a consumer in an Internet E-commerce platform; Screening out explicit behavior events based on the consumption behavior data, and counting the occurrence frequency, the occurrence time and the associated commodity category of each explicit behavior event to obtain explicit behavior characteristics reflecting the direct shopping behavior tendency of consumers; And screening out implicit behavior events based on the consumption behavior data, and analyzing attribute association relations between the implicit behavior events and commodity price intervals, types and brands to obtain implicit preference characteristics reflecting potential interests of consumers.
  3. 3. The method of claim 1, wherein performing behavior node association on the explicit behavior feature and the implicit preference feature in different context scenarios to obtain a node preference index of consumer behavior in different context scenarios specifically comprises: extracting behavior nodes corresponding to consumer behaviors under each context scene; Determining the association strength of each behavior node according to the explicit behavior characteristics and the implicit preference characteristics; And determining the node preference index of the consumption behavior of the consumer under different context scenes according to the association strength of each behavior node and the consumption behavior occurrence frequency of the corresponding behavior node.
  4. 4. The method of claim 1, wherein generating a short-term intent chain and a long-term preference chain for a consumer based on all node preference indices and the temporal order corresponding to each context scenario comprises: Dividing a consumer's consumption behavior time range into a short-term time window and a long-term time window; Generating a short-term intent chain of the consumer according to the node preference index in the short-term time window and the short-term intent trend of the consumer; a long-term preference chain for the consumer is generated from the node preference index and the long-term preference trend for the consumer over the long-term time window.
  5. 5. The method of claim 1, wherein extracting preference weights for each preference node in the element feature map under different context scenarios specifically comprises: determining influence factors of preference weights according to the occurrence frequency of preference nodes in each context scene, the association times of the preference nodes with the final purchasing behavior and the node preference indexes in the short-term intention chain and the long-term preference chain; Setting a demand weight coefficient for an influence factor according to actual application demands, and further carrying out preference scoring on each preference node according to the influence factor and the demand weight coefficient; and determining the preference weights of the preference nodes under different context scenes according to the preference scoring result.
  6. 6. The method of claim 1, wherein detecting the degree of matching of the element feature profile to the current consumer behavior feature comprises: Extracting target preference nodes of consumers in an Internet E-commerce platform and corresponding context scene information thereof based on the current consumption behavior characteristics; Selecting a preference node consistent with the context scene information from the element feature map, and calculating preference similarity between a target preference node and the selected preference node based on node attribute similarity and preference weight similarity; and determining the matching degree of the element feature map and the current consumption behavior feature according to all the preference similarity.
  7. 7. The method of claim 1, wherein updating the element feature map based on the preference weights and the current consumption behavior features corresponding to the preference nodes, and further generating the commodity recommendation list based on the updated element feature map specifically comprises: according to the behavior characteristics of the target preference node, the preference weight of the target preference node in the corresponding context scene in the element characteristic map is improved, and meanwhile, the preference weight of the associated node is synchronously adjusted; Correcting node attribute information and association strength between nodes in the element feature map based on the updating result of the preference weight to obtain an updated element feature map; and screening a plurality of nodes with top preference weights from the updated element feature patterns, acquiring corresponding commodity information, and sequencing the commodity information according to the preference weights from high to low to generate a commodity recommendation list for the current consumer.
  8. 8. A consumer behavior-based commodity recommendation system, comprising: the acquisition module is used for acquiring consumption behavior data of a consumer in the Internet E-commerce platform, and further extracting explicit behavior characteristics and implicit preference characteristics of the consumer in different context scenes; the processing module is used for carrying out behavior node association on the explicit behavior characteristics and the implicit preference characteristics under different context scenes to obtain node preference indexes of the consumption behaviors of the consumer under different context scenes; The processing module is further used for generating a short-term intention chain and a long-term preference chain of the consumer according to all node preference indexes and the time sequence corresponding to each context scene, and further constructing an element feature map reflecting consumer consumption behavior preference according to the mapping association relationship between the short-term intention chain and the long-term preference chain; the processing module is also used for extracting preference weights of the preference nodes in the element feature map under different context scenes; And the execution module is used for detecting the matching degree of the element feature pattern and the current consumption behavior feature, and when the matching degree is lower than a preset threshold value, the element feature pattern is subjected to preference updating based on the preference weight corresponding to each preference node and the current consumption behavior feature, and then a commodity recommendation list is generated based on the updated element feature pattern.
  9. 9. A computer device comprising a memory storing code and a processor, wherein the processor is configured to obtain the code and to perform the consumer behavior-based commodity recommendation method according to any one of claims 1 to 7.
  10. 10. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the consumer behavior based commodity recommendation method according to any one of claims 1 to 7.

Description

Commodity recommendation method and system based on consumer behaviors Technical Field The application relates to the technical field of internet advertisement services, in particular to a commodity recommendation method and system based on consumer behaviors. Background Under the background of rapid development of the Internet, an electronic commerce platform becomes an important channel for consumers to acquire commodities and services, as the scale of users is continuously enlarged and the types of commodities are increasingly rich, the platform needs to improve user experience and conversion efficiency through accurate recommendation services, personalized recommendation can help consumers to quickly find interesting commodities, can optimize advertisement putting effects, improve platform benefits, reduce information redundancy, and enable the consumers to obtain efficient and accurate shopping guidance in complex commodity environments, so that constructing an efficient and accurate recommendation system becomes an important means for improving competitiveness of the electronic commerce platform. In the prior art, the conventional commodity recommendation technology relies on historical clicking, purchasing or browsing behaviors of users to perform collaborative filtering or content-based recommendation, but these methods have obvious limitations in processing personalized requirements under different context scenes, specifically, the conventional methods are difficult to comprehensively integrate short-term interests and long-term preferences of users, and also cannot accurately describe consumption behavior characteristics of users under different scenes, so that recommendation results lack pertinence and instantaneity, for example, when the interest preferences of users under working scenes and leisure scenes have obvious differences, the conventional recommendation methods generally cannot dynamically adjust recommendation strategies, mismatching of recommendation or repeated commodity conditions easily occur, and user satisfaction is reduced, so how to realize personalized commodity recommendation based on different context scenes in an internet electronic commerce platform becomes a problem faced by the industry. Disclosure of Invention The application provides a commodity recommendation method and a commodity recommendation system based on consumer behaviors, which can realize commodity personalized recommendation based on different context scenes in an Internet E-commerce platform. In a first aspect, the present application provides a commodity recommendation method based on consumer behavior, comprising the steps of: Collecting consumption behavior data of a consumer in an Internet e-commerce platform, and further extracting explicit behavior characteristics and implicit preference characteristics of the consumer in different context scenes; Performing behavior node association on the explicit behavior features and the implicit preference features under different context scenes to obtain node preference indexes of the consumption behaviors of the consumer under different context scenes; Generating a short-term intention chain and a long-term preference chain of a consumer according to all node preference indexes and the time sequence corresponding to each context scene, and further constructing an element feature map reflecting consumer consumption behavior preference according to the mapping incidence relation between the short-term intention chain and the long-term preference chain; extracting preference weights of each preference node in the element feature map under different context scenes; Detecting the matching degree of the element feature map and the current consumption behavior feature, and when the matching degree is lower than a preset threshold value, updating the element feature map in a preference mode based on the preference weight corresponding to each preference node and the current consumption behavior feature, and further generating a commodity recommendation list based on the updated element feature map. Preferably, extracting the explicit behavior feature and the implicit preference feature of the consumer in different context scenes specifically includes: acquiring consumption behavior data of a consumer in an Internet E-commerce platform; Screening out explicit behavior events based on the consumption behavior data, and counting the occurrence frequency, the occurrence time and the associated commodity category of each explicit behavior event to obtain explicit behavior characteristics reflecting the direct shopping behavior tendency of consumers; And screening out implicit behavior events based on the consumption behavior data, and analyzing attribute association relations between the implicit behavior events and commodity price intervals, types and brands to obtain implicit preference characteristics reflecting potential interests of consumers. Preferably, performing