投資決策之困難在於投資時點的掌握，本研究整合不同的技術指標，使用適應性類神經模糊推論系統，建構一個預測模式系統，來預測股票投資買賣的時點。除了以預測準確率來評估模式系統的預測能力外，本研究配合「全數買賣」與「買入持有」兩種策略模擬交易，進一步以股資報酬率來評估使用預測模式系統之投資績效。研究資料取自某電腦公司，從民國86年1月1日至民國87年12月31日的股票資料。研究結果顯示，本研究之預測模式系統可以有效地預測股票的買賣時點。

The major difficulty of investment is to determine the proper timing of transactions. This research applies a technical analysis method together with an adaptive neuro-fuzzy inference system to establish a model for predicting buying/holding/selling signals of stock investment. We use the predicting rate and the rate of return to evaluate the model’s performance. The data are collected form an electronic company in Taiwan during the period form 1/1/1997 to 12/31/1998. The results indicate that the adaptive neuro-fuzzy model can help to predict the proper timing for investment and get better return on investment.