Mining Of The Financial Micro-Blog Emotion Attitude Based On Dictionary

In recent years, with the rapid development of Chinese economy, Chinese stock market develops rapidly.So far, China stock market has had2467listing Corporations, and the total market capitalization of the Shanghai and Shenzhen stock market isf$23.5.The mainland stock market has become the third largest market in the world. There are about160million Chinese investors in stock market,and they are interested in Internet financial news.As a new social networking tool,Micro-blog has become the domestic second big social networking media.Because of its brief writing,convenient,real-time interactive, it is so popular.It is also the second largest source of public opinion, it seems that all the events happened in our country will effectively spread in micro-blog. It is necessary to analyse sentiment in financial micro-blog,because it can predict the market trend,and provide decision-making basis for the financial industry employees and investors.Aimed at emotional attitude in the micro-blog of finance and economics analysis study, I build a complete classification model, mainly from the standardization, classification, named entity recognition, emotion research analysis, trend prediction, etc. But in the concrete process, I focus on the emotional analysis.Classification of emotional tendency is also known as Opinion Mining or emotional polarity classification.It can be understood as a user of a certain object to express their own views held by the attitude for agree or against, neutrality, which is often said that the positive emotion, negative emotion and neutral emotion.In the concrete implementation process of the theis,the main content includes the following sections:(1)the full name and abbreviationof company are studied and combined with the financial sector’s unique emotion word, use emotional tendency of mutual information algorithm (SO-PMI) to build the financial domain dictionary.(2)analyze the characteristics of Chinese weibo, in combining with the characteristics of network language and the financial language to build the network language dictionary and negativity, adverbs of degree and emoticons dictionary.(3) emotional weighted method is proposed, it can be applied to build all kinds of dictionary emotion classificationand it realise quantitative calculation on classification of emotional valueAt last, through sina API for a period to get microblog news which contain the name of the company.After preprocessing, segmentation and feature selection, I use a dictionary of emotion classification method for classification.The result indicates the importance of the Financial sector dictionary,network dictionary and expressiondictionary.Every dictionary under the complete experimental data were compared with the actual market trends.The results show that the experimental data has a realistic significance in the real life, through further study it can be applied in stock investment.