A One-Class Collaborative Filtering Method Fused with Personality Traits and Item Labels

Abstract

The invention discloses a single-category collaborative filtering method that combines personality traits and item tags, including: firstly calculating the similarity between users based on user personality traits, and calculating the user's preference for items based on the similarity; then calculating based on the item tags The degree of user preference for items; followed by the fusion of personality-based user preference for items and user preference for items based on item tags to obtain the total preference degree of users for items, and use the total preference degree to construct a matrix decomposition model; finally, according to the model Make recommendations. The present invention uses personality traits and item labels as additional information to identify positive and negative examples mixed together in missing data, thereby effectively solving the data sparse problem and the cold start problem of new users, and then improving the accuracy of personalized recommendation.

Description

technical field

[0001] The invention belongs to the field of electronic commerce, in particular to a matrix decomposition method (PTMF) combining personality traits and item labels. Background technique

[0002] With the development of information technology, the expansion of information resources and the rapid development of e-commerce, it has become a difficult and expensive obstacle for users to find product information they are interested in; Effectively improving the purchase rate of users has become their primary consideration. Recommender systems can overcome this obstacle by providing users with personalized items, products or services that meet user needs. Collaborative filtering technology is one of the earliest and most successful technologies for personalized recommendation applications. It can provide technical support for users' purchase decisions based on the similarity between items or users. Collaborative filtering has been extended and practically applied...

Examples

Embodiment Construction

[0042] In this example, if figure 1 As shown, a single-category collaborative filtering method that combines personality traits and item tags is carried out in the following steps:

[0043] Step 1. Use the two-dimensional table R={u,i} to represent the user’s behavior records on items, use the two-dimensional table P={u,p} to represent the user’s personality trait data, and use the two-dimensional table T={i,tag } represents the tag data of the item; among them, u={u 1 ,u 2 ,...,u n ,...,u |N|} represents the set of users, u n Indicates the nth user, n=1,2,...,|N|, |N| indicates the total number of users; i={i 1 ,i 2 ,...,i m ,...,i |M|} represents a collection of items, i m Indicates the mth item, m=1,2,..., |M|, |M| indicates the total number of items; p={p 1 ,p 2 ,...,p n ,...,p |N|}Represents the information set of user personality traits, p n Indicates the nth user u n personality traits and have: Indicates the nth user u n The jth personality trait of ...

technical field

[0001] The invention belongs to the field of electronic commerce, in particular to a matrix decomposition method (PTMF) combining personality traits and item labels. Background technique

[0002] With the development of information technology, the expansion of information resources and the rapid development of e-commerce, it has become a difficult and expensive obstacle for users to find product information they are interested in; Effectively improving the purchase rate of users has become their primary consideration. Recommender systems can overcome this obstacle by providing users with personalized items, products or services that meet user needs. Collaborative filtering technology is one of the earliest and most successful technologies for personalized recommendation applications. It can provide technical support for users' purchase decisions based on the similarity between items or users. Collaborative filtering has been extended and practically applied...

Examples

Embodiment Construction

[0042] In this example, if figure 1 As shown, a single-category collaborative filtering method that combines personality traits and item tags is carried out in the following steps:

[0043] Step 1. Use the two-dimensional table R={u,i} to represent the user’s behavior records on items, use the two-dimensional table P={u,p} to represent the user’s personality trait data, and use the two-dimensional table T={i,tag } represents the tag data of the item; among them, u={u 1 ,u 2 ,...,u n ,...,u |N|} represents the set of users, u n Indicates the nth user, n=1,2,...,|N|, |N| indicates the total number of users; i={i 1 ,i 2 ,...,i m ,...,i |M|} represents a collection of items, i m Indicates the mth item, m=1,2,..., |M|, |M| indicates the total number of items; p={p 1 ,p 2 ,...,p n ,...,p |N|}Represents the information set of user personality traits, p n Indicates the nth user u n personality traits and have: Indicates the nth user u n The jth personality trait of ...

Images