Research On Commodity Recommendation System Of A Company's Network Shopping Guide

With the rapid development of e-commerce industry for more than ten years,consumers' demands for good services of online sales are becoming stronger and stronger.But the general B2C website guides cannot satisfy consumers' demands,such as facing online homogeneous commodities,they cannot provide reasonable purchase opinions.A company in line with the purpose of better customer service,has established a network shopping guide platform to provide scientific consumer advice to customers who have difficulty in choosing.In the actual operation of the platform,as more and more goods are distributed in the commodity pool,the online shopping guide becomes time-consuming and laborious,and the conversion rate of goods recommended to consumers is not high,which seriously affects the activity of online shopping guide.In order to improve the activity of online shopping guide,this paper decided to build a recommendation system of online shopping guide to help online shopping guide select products in the way of recommendation,which is of great significance for A company to improve customer service and performance.But in the process of building the system,there are some problems,such as sparsity,cold start problem,low recommendation accuracy and efficiency.Aiming at the low recommendation accuracy and efficiency of the system,from the point of view of consumers and online guided purchasing,this paper first classifies commodities into positive feedback and negative feedback commodities,eliminates negative feedback commodities(commodities that consumers do not like);then,emotional intensity of positive feedback commodities is quantified by emotional analysis of their evaluation eigenvalues.Finally,according to the score,emotional intensity and commission amount,the comprehensive score of the goods is calculated.Through the above processing,not only can the recommended pool be the goods that consumers like,but also can get reasonable commission for online shopping guide,which indirectly improves the accuracy and efficiency of the algorithm.To solve the problem of data sparsity,firstly,the correlation degree is introduced to calculate the correlation matrix,then the candidate item sets are obtained according to the correlation matrix,and finally,the network guided purchase prediction score and recommendation are generated by combining the similarity of items.In order to solve the cold start problem,we first extract the characteristic value of Network Guided purchasing,obtain the obvious characteristics of Network Guided purchasing,and calculate the similarity of Network Guided purchasing characteristics.Finally,we construct the recommendation scheme based on the similarity of Network Guided purchasing characteristics.Based on the model of recommendation algorithm of A company's network-guided purchase,this paper completes the requirement analysis and design of the recommendation system of A company's network-guided purchase,and completes the development of the system.The test results show that the product recommendation system of A company's Network Guided purchase basically realizes the functional requirements of the design.This paper establishes a recommendation system for online guided purchasing of Company A,which can help online guided purchasing select the right goods,thus improving the activity of online guided purchasing.At the same time,the research of this paper can provide some reference for the online shopping guide platform in building the commodity recommendation system.This paper contains 25 figures,15 tables and 40 references.