Research On Text Sentiment Analysis Technology

In recent years, with the development of the Internet technology and the proposition and application of many new ideas, such as the blogs,forums,customer service flatform,e-commerce platform, we are marching into a new era of information. Network which continuously changes people’s traditional life, work and study style has become a closely part of the development of all trades and professions and affection of people’s life. For a large number of information text enriching emotional color, but the manifestations of these words is not only the textual representation, there will be pictures, animations, and other manifestations of this unstructured.we face processing cycles and higher cost if only by artificial.There is an urgent need for people to look for a method to draw required emotional information from the large number of information text promptly, and is able to carry on the emotional sentiment analysis. Under this demand background, the text sentiment analysis research arised.As new areas are emerging, the amount of information in various fields is rapidly increasing,due to strong a word no longer have a strong tendency in the test domain, Under this demand background, interdisciplinary analysis research arised. Interdisciplinary sentiment analysis is a relatively full new research field; the present research work is lesser. For text is composed of words, emotional dictionary building is the basic content of the text sentiment analysis, the article focuses on that on the basis of constructing interdisciplinary dictionary algorithm, with the goal of the interdisciplinary text sentiment analysis, performing research on the following aspects:First, the article introduces background and current situation of the text sentiment analysis technology, and on the basis of analyzing current situation of the emotional sentiment analysis home and abroad, analyzes construction of emotional dictionary and challenges the text sentiment analysis faces, and finally further expressly states the significance and necessity of the interdisciplinary text sentiment analysis research.Second, the article highlights construction technology of text sentiment analysis research. The article introduces the process of text orientation analysis.fristly, Get the key words which can reflect the emotions of the document by the technology of Chinese wore segmentation; secondly, according to the dictionary construction algorithm to build the dictionary. This chapter describes the construction technology of the interdisciplinary emotional dictionary and generic emotional dictionary. Generally speaking, Universal Dictionary building technology, including building generic dictionary based on HowNet, based on function optimization method and optimization method based on Modularity; a extended information bottleneck-based construction method of the interdisciplinary dictionary is presented.This chapter from the perspective of both the theoretical and experimental analysis based on improved information bottleneck algorithm characteristics of the convergence and effectiveness of the method. Finally, Text Orientation measure the inclination of the first use of the computing word similarity method to get the word, according to the principle of cumulative, and then the inclination of the statement and then the inclination of the text.Third, the article main introduce the interdisciplinary text sentiment analysis technology. Aiming at the interdisciplinary emotional sentiment analysis research, the article states the field transplantation of the application technology in the general text sentiment analysis. In order to avoid the passive impact of eigen space difference on the performance of classifier, by analyzing the common and the difference between eigen spaces of the source field and the objection field, the field transplantation model of supervised learning the sentiment analysis method is put forward; in order to balance data in the source field and acquire information of objection field, the bayesian learning-based emotional transplantation model is put forward; the theme idea of the figure sorting model-based interdisciplinary sentiment analysis algorithm is on the basis of the figure sorting algorithm, performing iteration on the sentiment analysis with the labeled accurate label in training texts in the source field and the initialized false label computed in test text in the objection field.Fourth, the article mainly introduces experimental data sets and experimental evaluation standard and experimental platform that the text sentiment analysis employs.