| Summary: | 碩士 === 銘傳大學 === 資訊工程學系碩士班 === 105 === Stock market is a business differs from general profit organizations which providing needs for economic development and also undertaking the role of financial intermediation by collecting funds from the investors. Each country takes information disclosure as an important way of administrating its own stock market. Such as in Taiwan, the investors could inquire about the financial and non-financial information from the platform of MOPS（Market Observation Post System）, investors in the U.S. could do this by logging onto the EDGAR（Electronic Data Gathering, Analysis, and Retrieval System） under Securities and Exchange Commission. Nevertheless, does it really help the investors to make policies by offering huge amount of information? Will there exist some risks by facing overloaded information? Related stock trading laws in Taiwan request all who take responsibilities should offer to disclosure the information which are needed, thus the participators in the market are expected to make the correct investing policies by efficiently using these information to fit to the cost benefit of all the investments in accomplishing these information.
This research will use the techniqueof documents summarization in text mining to gather the differentiated key words of anticipated companies which will leave public from thematerial information onMarket Observation Post System. Secondly, it will establish four kinds of classifiers based on eigenvector either to predict the value of probability or whether or not a company will leave public to obtain the latent risks in the market to offer the investors early-warning they need.
|
|---|
