Text Texture Modeland Its Application

With the development of the Internet, information on the Internet hasexperienced a dynamite increasement. Nowadays the Internet has becomeone of the most important vehicle for people to express their views andcomments. These information implies a high value to be dig up. Now themost popular social networks, e-commerce has become the focus ofpeople’s attention. their information can be quite a true reflection of publicattitudes toward an event or a commodity. Research on text emotionaltendency will help people find out public opinion tendency, the value ofgoods and so on. Before the text emotional tendency analysis, we have toexclude the interference of non-emotional texts. That is to say we have totell subjective texts from objective texts. Therefore, this paper willresearch on the emotional tendency analysis and subjective and objectivetext classification.This paper aim to find out a differentiated approach with traditionaltext classification, focusing on sentence structure, sentence pattern,words’ dependencies. This paper will also combine text texture withtraditional classification which based on the emotional words and phrasesand subjectivity classification which based on semantic rules to get betterresults.The main work of this paper include:1. It summarizes the entireprocess of a classic text classification system, and then compare the existing common technique of text classification--text representation,feature selection, features’ weight calculation, text classification andevaluation on text classification results. It also introduces a variety oftraditional text orientation classification methods and text subjectivity andobjectivity analysis methods. It also present their respective advantagesand limits.2. It gives out a detailed description of text texture, andpropose several application scenarios based on the basics of textclassification system. Text textures used in the experiment are: sentenceparallelism, negative modification, adverb modification, turning sentenceand so on.3. It focuses on the subjectivity and objectivity classification,introduces text texture into subjectivity and objectivity classification in aninnovative way. It draws the system block diagram, which includes wordbreaker, Stanford Parser, features extraction, feature construction, classifier,classification results evaluator. After that it describes how to implementeach part of them. Finally it gives the comparison with the existingclassification, which includes experimental procedures, experiment data,analysis and evaluation.4. It focuses on the emotional orientationidentification, and also introduces a series of text texture features intotraining and classification. It draw the system block diagram of theclassification process. Finally, it gives the comparison with the existingclassification experimental procedures, data, analysis and evaluation.5. Itsummarizes the experimental results in subjectivity classification andemotional orientation identification, and the prospect of its application inthe real world.The research results can be applied to the network media, emotion recognition of product reviews, emotion recognition of literary works,emotion recognition of movie reviews and so on.