Analysis Of Weibo Sentiment Orientation Based On Contextual View

Micro-blog is widely used as a new social media.The number of its data exchange everyday grows in an explosive speed,which provides a good corpus environment for Natural Language Processing.The micro-blog text possesses the same features as the short text that they both have short length,strong emotion expression and single topic,which needs to be processed by sentiment analysis methods different from the long texts.The text aims to improve the sentiment analysis in short texts,put the context information into affective computing,establish micro-blog emotion dictionary for micro-blogging smiley and eventually get the result of emotional tendency.To begin with,this paper studies and explores the processing technology for micro-blog review text;then it constructs the micro-blog emotion dictionary by the means of fusing the micro-blogging smiley into it automatically and extracts the context information from micro-blog comments and finally integrates the above information to get the get the result of emotional tendency.The paper mainly includes the following three aspects:(1)This paper puts forward the automatic construction method of the emotion dictionary based on micro-blogging smiley and applies it to the sentiment analysis of micro-blogMicro-blog comment text usually has a subjective emotion and is widely used in various areas,thus the emotional information of micro-blog comments should be correctly analyzed before the sentiment analysis is conducted.The degree of the perfection of the emotional dictionary determines the level of the accuracy of the sentiment analysis.The author of the thesis puts forward the automatic construction method of the emotion dictionary based on the lexicon construction and the influence of the micro-blogging smiley on the emotional tendency of micro blog.Firstly,the method employs the COAE2015 task—a provided training data to classify the words of the micro-blog comment text and cancel the repetitive information to gain the emotional words and obtain their occurrence times in the positive and negative corpus;then calculates the size of the emotional tendency by use of PMI to generate emotional dictionary at last.(2)This paper studies the influence of contextual view on emotional predisposition,and proposes the analysis method of emotion tendency computing based on context informationIt is necessary to define the context information in texts before the analysis of emotion tendency computing.Users are more or less influence by the original texts and previous comments when they make comments.Considering that,the author proposes the analysis method of emotion tendency computing based on context information,which is applied to the emotional analysis of micro-blog comment text.Firstly,the comment text is numbered in term of the sequence of making comments.Then the comment text is made weigh in line with the number.Finally,the emotional tendency of the algorithm is described based on the results from established emotional dictionary.(3)This paper designs the prototype system of micro-blog emotional tendency computing based on the context information and puts it into use.In the process of analyzing the emotional tendency of micro-blog comment text,corresponding functional modules for every aspect are designed and the prototype system of micro-blog emotional tendency computing based on the context information comes into use.This system can extract and analyze the information in comment text from micro-blog.Ultimately,it provides intuitive emotional tendency for users.