Data Mining In Finance By Discussing IPO Underpricing

Due to the rapid advance of information technology, data storage is not as difficult as before. The amount of data recorded by human keeps on increasing continuously. How to use these data is the study direction of Data mining. Data mining can help us integrate, generalize and analyzing large amounts of information.Over the years, economists study IPO in different ways and for different purposes. They summarize a number of hypothesis of IPO underpricing, most of which are the explanations of one existing phenomenon. If we study these factors as a whole, what the conclusion will be? How about the relationship between them? Can we use historical IPO data to predict the underpricing level of future IPO? In the past due to limitations of information technology, these issues are difficult to analysis. With the development of information technology, we are able to analysis massive data sets. Using data mining method, we can explore the"unknown knowledge"which is behind huge data sets.The paper reviews the history of china stock IPO market, and summarizes the conclusion of relevant thesis. We gather the factors that will affect IPO underpricing. Using data mining method, we show IPO data in different ways. After that, we use several data mining modeling tools to analysis the impact of these factors. According the accuracy of the models, we get the most reliable model, which show the importance of every factor. This paper explains the study direction of this issue and gives some suggestions to those who apply data mining method in finance area.