Document Type



In Taiwan stock market, it has been accumulated large amounts of time series stock data and successful investment strategies. The stock price, which is impacted by various factors, is the result of buyer-seller investment strategies. Since the stock price reflects numerous factors, its pattern can be described as the strategies of investors.

In this paper, pattern recognition concept is adapted to match the current stock price trend with the repeatedly appearing past price data. Accordingly, a new method is introduced in this research that extracting features quickly from stock time series chart to find out the most critical feature points. The matching can be processed via the corresponding information of the feature points. In other words, the goal is to seek for the historical repeatedly appearing patterns, namely the similar trend, offering the investors to make investment strategies.