1. Background[1]
Investment decisions are about making choice. Investors need to decide what asset to be invested. This is an important decision because these assets are the means by which investors transfer today’s purchasing power to the future. In effect, investor must decide on a portfolio of asset to own. A portfolio is simply a combination of assets designed to serve as store of value. Poor management of these assets may destroy the portfolio’s value, and investor will then not achieve their financial goals.
There are many assets like stocks, bonds, derivatives that investors may include in the portfolio. In Hong Kong, stock portfolio is the most common investment. But what stocks have higher return? What stocks are risky? What stocks in the portfolio that it has higher return? Many investors may use fundamental analysis to analysis financial data for answering above questions. In the last decade, some researches applied data mining techniques on financial market. Data mining is the process of automatically discovery useful information in large data repositories. It can be used to support a wide range of business intelligence applications such as customer profiling, targeted marketing, store layout.
2. Motivation
In America, there are some research papers[2] about applying clustering technique on America financial market. For example, using Self-organizing map(SOM) to cluster stocks and financial ratio for fundamental analysis, using SOM to find the valuable stock. These all researches want to find the characteristic of the stocks.
However, most these researches use SOM clustering technique, and focus on America financial market. It seems that few researches do similar thing in Hong Kong stock market. Moreover, these researches only want to find the characteristic of stocks. In real case, investors will not only purchase one stocks. They will construct a stock portfolio to eliminate the risk.
So I want to apply