Commodity similarity calculation method and commodity recommendation system based on image similarity

Abstract

The invention relates to the field of internet electronic commerce, in particular to a commodity similarity calculation method and a commodity recommending system based on image similarity. The method includes: preprocessing a target image, to be specific, removing image differences caused by changes in light conditions such as brightness and chromatic aberration; processing the target image to detect a foreground frame; converting a community image in the foreground frame into pixel images different in scale by means of bilinear interpolation, and acquiring attribute features, in different dimensions, of the commodity image in the foreground frame under different scales; calculating attribute feature similarities, under different scales, between an attribute feature vector of the commodity image in the foreground frame and an attribute feature vector of a commodity sample image; according to a decision forest model and the attribute feature similarities under different scales, calculating commodity image similarities, under the pixel images of different scales, between the commodity image in the foreground frame and the commodity sample image; using the commodity image as a uniform identifier of a commodity on different commercial platforms. The commodity similarity calculation method and the commodity recommending system have the advantage that reliability of the system is greatly improved.

Description

technical field [0001] The invention relates to the technical field of Internet e-commerce, in particular to a product similarity calculation method and a product recommendation system based on image similarity. technical background [0002] Collaborative filtering is a widely used technique in current user recommendation. Collaborative filtering analyzes user interests, finds similar (interested) users of a specified user in the user group, and synthesizes similar users' evaluations of a certain information to form a prediction of the specified user's preference for this information. [0003] Collaborative filtering establishes a user-product matrix through the user's purchase behavior, and based on this matrix, "collaborative filtering of similar users" and "collaborative filtering of related products" are performed. [0004] "Collaborative filtering of similar users" starts from the "user row" of the user-product matrix, and obtains the similarity of user purchase behavi...

AI Technical Summary

Problems solved by technology

[0013] 1) The user's historical purchase data only comes from the purchase behavior of an e-commerce system, and the limited historical purchase data leads to the sparsity of the user-product matrix;

[0014] As far as the "user row" of the matrix is ​​concerned, it limits the ability to describe the user's purchase intention, which in turn limits the accuracy of judgment of similar users and affects the quality of recommendations; especially for new users of the platform, the existing system cannot recommend

[0015] As far as the "product column" of the matrix is ​​concerned, it reduces the probability that related products occur in multiple different user purchase behaviors, reduces the accuracy of judgment of related products, and affects the quality of recommendations

[0016] 2) The same product may have multiple different identifiers, and the same product from different merchants is identified as different products in the user-product matrix of collaborative filtering, which increases the sparsity of the user-product matrix

[0017] 3) Whether it is "collaborative filtering of similar users" or "collaborative filtering of related products", the recommended products based on it must come from the e-commerce system currently browsed by the user, and products purchased in other e-commerce systems or offline cannot as a recommended candidate

Images