Research And Analysis Methods In Identifing Opinion Leaders In Stock Network Forum

With the rapid development of the Internet, people entered the information explosion period from the meager information period. BBS has become an important information platform for its simplicity、convenience and speediness. Since Chinese reform opening policy for more than 30 years, its economy has risen rapidly, in the same time, people accumulate lots of material wealth. With the great enthusiasm of public investment, stock market gets a lot of investors for its simplicity and low capital demand. Among these investors, opinion leaders are those who have made deep market research, their opinions can be accepted and followed by other investors.Beginning with the intrinsic characteristics of opinion leaders, and based on the research of text structure and data grab method of Eastmoney guba, we design and complete a data grab and analysis system for Eastmoney guba. The main studies are identification methods of the opinion leaders with cluster analysis, as well as their performances in different market conditions.Main contents include:(1)For great capacity comments of Eastmoney guba, we design an automatic data grab and extraction system based on the research of the data crawler technology, guba URL variation and the character structure. This system can grab data in guba exactly.(2)We put forward a two-step cluster method to pick up opinion leaders in the forum. It begins with the common characteristic of the opinion leaders, and calculates the eigenvalue within each time period. Through this two-step cluster method, the results show that it can be effectively pick up opinion leaders.(3)Combing with the variant time period, though the data analysis of different market conditions in the forum, we explore the opinion leader’s performances and regulations in the forum.