Weighted voting-based e-commerce platform user recommendation method

Abstract

The invention discloses an e-commerce platform user recommendation method based on a weighted voting mode. The method comprises the following steps: 1, converting unstructured information of e-commerce platform commodities into a structured data format; 2, designing a recommendation model based on commodity content similarity; 3, designing a recommendation model based on user historical search behavior similarity; 4, designing an e-commerce platform user recommendation model based on a weighted voting mode; 5, embedding a weighted voting-based e-commerce platform user recommendation method into the data processing module, and recommending commodities to the user in real time by an e-commerce platform user recommendation system according to the actually measured data. In order to solve the problem of e-commerce platform user recommendation, a recommendation model based on commodity content similarity and a recommendation model based on user historical search behavior similarity are fused according to voting vector decision. According to the method, the relevance between commodities and the relevance of user historical search are fully mined, personalized recommended commodity labels are generated, and recommendation of e-commerce platform products is completed.

Description

Technical field [0001] The present invention relates to the field of e-commerce platforms, in particular to the recommended methods for e-commerce platform based on weighted voting. Background technique [0002] The emerging technology has subverted the form of traditional living in people, especially in the shopping sector, significant market, the earliest market to subsequent large shops, large supermarkets, and further evolve into the current large-scale e-commerce, the quantity of goods from a few Hundreds of rapid increases to 10,000 kinds, and now there are more hundreds of millions of species, because the traditional shopping environment has made great changes, and the general people will not have enough energy and time than three in traditional shopping. [0003] In these years, the data retrieval of this type has been a new model in accordance with the customer's own needs, providing customers with a set of personalized information services in a huge database in a huge d...

AI Technical Summary

Problems solved by technology

Although some results have been achieved, there are still some remaining problems that need to be solved urgently, such as data sparsity, cold start, low recommendation accuracy, etc.

Images