Font Size: a A A

Property Comments Sentiment Analysis Based On The Theme Classification Feature

Posted on:2012-11-13Degree:MasterType:Thesis
Country:ChinaCandidate:C XuFull Text:PDF
GTID:2218330368992256Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the growing popularity of Internet, people use the Internet to obtain information in addition to the purpose, the expression of personal views, opinions, views also increasing. Determine the characteristics of the network's own people on the Internet more in line with the views expressed by their real thoughts. The current level of quality property management concerns owners increasingly, the traditional satisfaction survey did not reflect the views of the owners in the relevant forums, comment on the information of property owners can more truly reflect the level of satisfaction. Therefore analysis of the property has some practical significance comment.Sentiment analysis research goal is to identify the subjectivity of the text contained the sentence, and the sentimental tendency to judge, so you can analyze the existing theory of sentimental comments on the property information contained in satisfaction of judgments.In this paper, Kunshan Forum as the basic data sources, data collection on the forum, topic extraction and classification, the use of sentimental words to the core set of keyword matching approach to sentiment analysis, and ultimately the satisfaction of the formation of a web page for users to access the chart. Comment by property in this area, on the Classification of the dictionary and emotional word sets were optimized and the supplement to the feelings of more accurately reflect the conclusions of Property Management Satisfaction.Through systematic analysis of the results truly reflect the status of property management for owners to choose the living environment and property management companies to provide a reference.
Keywords/Search Tags:sentimental analysis, sentimental words set, Automatic segmentation, Classification feature
PDF Full Text Request
Related items