Product similarity judgment method and device and server cluster

Abstract

The invention provides a product similarity judgment method and device and a server cluster. According to the embodiment, when the similarity among products needs to be judged to recommend similar products to a user who is possibly interested in the similar products or realize classification management and the like of a large quantity of products according to the similarity, at least one piece ofproperty information of a product is obtained, wherein the property information refers to factors possibly influencing product selection by the user or factors influencing classification by the user,such as a CPU, a hard disk, a memory, size, thickness and other property information of a notebook computer; and next, a current property weight value corresponding to each piece of property information is determined, wherein the current property weight is a probability reflecting the influence of the corresponding property information on user fondness of the product with the corresponding property information. In this way, based on the property information and the current property weights corresponding to the property information, the similarity among multiple products can be calculated moreaccurately, and therefore the application reliability based on a product similarity calculation result is improved.

Description

technical field [0001] This application mainly relates to the field of similarity calculation, and more specifically relates to a product similarity judgment method, device and server cluster. Background technique [0002] With the development of big data technology, companies usually track users' usage of products to improve product design, adjust product business strategies, expand customer groups, etc., so that companies can gain more benefits and market competitive advantages. [0003] In practical applications, companies usually analyze the behavior information of users using different products and the characteristics of the groups to which users belong, and combine the analysis results to recommend similar products that users may be interested in, or implement product classification management. [0004] In the prior art, product similarity is usually determined by calculating the similarity of text information such as product names and descriptions. However, since info...

AI Technical Summary

Problems solved by technology

[0004] In the prior art, product similarity is usually determined by calculating the similarity of text information such as product names and descriptions. However, since information such as product names and descriptions often cannot truly reflect product information, this will affect product similarity. The accuracy of degree calculation, which in turn affects the reliability of applications such as product recommendation or product classification based on product similarity

Images