修正移動平均法於台灣股市之報酬率研究

Abstract

□劓吤郁〞k為過去股票市場上常見之交易法則。過去的研究發現移動平均法的投資績效不具有良好之效率性。本研究建立一修正式之移動平均法，運用資料探勘工具-決策樹的方法將報酬之買點做分類，根據分類的結果建立新的交易法則。經台灣悒囿煽□桮痕G發現，平均年複合報酬率由4.17%提昇至7.70%。

The MA (moving average) trading rule for stocks has been extensively studied in various stock markets. Previous research shows that the MA trading rule may not be widely effective. This research aims to develop a method for enhancing the MA trading rule. The basic idea is to hierarchically classify the transaction points produced by the MA rule into several groups, of which some are positive and the others are negative in the rate of return. The classification mechanism is implemented by decision tree—a data mining tool. Taiwan stock exchange market has been used to justify the performance of the enhanced MA trading rule. Experiment results show that the compound annual rate of return increases from 4.17% to 7.70%, by enhancing the MA trading rule.