Design And Implementation Of Linear Product Intelligent Shopping Guide System

With the continuous development of e-commerce and network services in recent years,e-commerce shopping for traditional industrial products has become a new type of shopping platform.Some manufacturers of linear products have also launched their own Taobao stores to keep up with the trend.However,there is a certain technical threshold for consumers to purchase linear pro ducts on e-commerce platforms like Taobao,which will greatly increase the knowledge cost of users.Therefore,we need a customized platform,and some customized recommendation services added to the system to bring user more professional experience.Currently,the existing linear product shopping guide system mostly adopts the traditional single architecture.The single architecture is increasingly prominent in the lack of scalability,which is easy to affect the business development when facing the constantly updated product information and the changing needs from the customers.Therefore,this thesis designs and implements an intelligent shopping guide system for linear product based on micro-service architecture.The main work includes:(1)Analyzing the demand and designing the intelligent shopping guide system for linear product.According to the results of demand analysis,system major functions are designed and split into multiple modules.The three main modules are designed accordingly and the system database is built.Finally,the system framework is setup according to the relevant design,and the shopping guide system is split accordingly by the principle of micro-service design.(2)A recommendation algorithm for linear product is designed.The recommendation algorithm starts from the two directions of collaborative-filtering-based recommendation and content-based recommendation and get them integrated together for the recommendation.Firstly,this thesis introduced the collaborative filtering algorithm which is normally used and its implementation.And then proposed an improved algorithm which integrated product features.This algorithm extracts product features through TF-IDF model and then solved the sparse score matrix and cold start problems of collaborative filtering algorithm by this score filling matrix.In addition,this thesis integrated both old and new factors into the linear product recommendation algorithm to make the recommended linear product time-sensitive and more in line with user needs.(3)The system is implemented,deployed,and tested.The system brings the features of product searching,online sales,product management,linear product recommendation,collocation recommendation,push without password protection etc.and passed the functional test and performance test.