Design And Implementation Of Commodity Evaluations Query System

This paper designs a classification method based on short text emotion classification model,and develops a commodity evaluations query system on this basis.This paper has completed the following work: First,this paper studied the vocabulary characteristics of short text,and designed an emotional model based on the emotional dictionary and the SO-PMI algorithm.Then,this paper proved that the emotion model performs well in text classification through contrast experiment.Second,this paper designed a commodity evaluations query system based on the emotional model.The commodity evaluations query system can help people to know a commodity quickly and conveniently.The system is divided into three modules: one is the crawler module,and a special reptilian program is written for the Jingdong,Taobao,the Pacific and other websites,and the evaluations is obtained from these websites regularly every day.Two is the classification procedure,it divides the evaluations into good and bad reviews.Three is the evaluations query system,to provide users with commodity evaluations query and contrast service.The workflow of the system is as follows,first,all original evaluations are obtained from the Jingdong,Taobao,the Pacific and other popular websites through web crawling,then them are divided into good and bad evaluation by the emotional model designed in this paper.At last,through the evaluations query system,people can search a certain commodity's positive evaluations and negative evaluations,what's more people can also compare two commodities' evaluations.During development,Java is used as programming language,MySQL is used as database,SSH2 framework,Lucene,Ajax and other technologies are used to complete the commodity evaluations query system.After testing,the commodity evaluations query system can classify the evaluation correctly and people can easily use the query system.