A study of stock investment based on multilayer intelligent selection system and Markowitz theory

• Main Authors: Rong-Tai Liang, 梁榮泰
• Other Authors: Ting-Cheng Chang
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/38806500712845736941

碩士 === 嶺東科技大學 === 財務金融研究所 === 96 === This thesis proposes a GVPRS-model to train and filter aim at financial data of listed company, more attempt find out by company that worth being invested in order to instead of traditionally artificial methods to select. And a different one is, the definition of worth being invested is not good performance stocks or leading stock that general investors like, but define the company as potential company. The model of this research has combined the (GRS-model), VPRS-Model, K-means clustering, Grey relational analysis and Neural-fuzzy theory to train and cluster the financial data of listed company to merge decision attribute and delete condition attribute, and then find out the company that NEGn change into POSp, finally we use Grey Relational Grade of Grey system theory to rank the company that has been filtered as judging the standard worth being invested. After filtering, we try to simulate the portfolio and cluster the company by traditional Rough Set Theory, decide when to buy in this way, predict future tendency by Grey prediction and decide stop limit point in this way. In addition, Markowtiz, grey relation and equal weight are used separately to be an allocation method in investment, and the result is compared. It can be fund that Markowitz is better than the other method in empirical result.