Risk Analysis Of GEM Stock Prospecting Based On Text Mining Technology

October 30, 2009, China Growth Enterprise Market had been opened in Shenzhen finally after a long time prepareation.The opening of China Growth Enterprise Market(GEM) made a way to equity trading for many medium and small-sized enterprises which can not get into the main board market. Moreover, since the GEM companies focused on the growth of the companies,they often face higher risks. Therefore, this paper intends to analyst the disclose risk information in prospectus of the listed companies on the GEM. This paper study the prospectus mainly because of the prospectus is one of the most important information disclosure with the legal literature of listed companies, they have important reference value for investors and regulators.Firstly, this paper make quantitative analysis for different enterprises prospectus Risk Information section, get the descriptive statistics of the amount of information disclosed and the length the risk factors disclosure, then comparing information from different areas of enterprise risk disclosures.When analyst the contents of the prospectus text, this paper introduces the text mining technology. At first, I do some reduction of the noise and other pretreatment to the prospectus text data, and utilization of the Institute of Computing Technology Research Institute of Chinese Lexical Analysis System--ICTCLAS, text data, word processing. Transform the unstructured data into structured mode, then reduce dimensionality of text data.When enterprises disclose the risk, it take turns in accordance with the importance of affecting the performance. So this paper extract primary risks of various enterprises, for which the key words for analysis. The programming TF-IDF algorithm use word frequency TF to show the commonality of enterprises risk in different regions, use TF-IDF values to extracted specific risk from different companies. Combining with the distribution of corporate-owned industries in the region, summary of the specific risk to be listed on the GEM corporate disclosure presents its economic characteristics in the region.By studying the risks of the GEM listing prospectus, we draw the status of China’s GEM companies risk information disclosed in the prospectus, and improve the disclosure of thinking and put forward suggestions. This article applies Internet traditional text mining techniques to the analysis of traditional prospectus, however the mining model established in this paper is not perfect. After the study, I will try to combine the preprocessing, segmentation, feature extraction and other mining modules together to form analysis process for mining prospectus information in the future.