Kahneman and Tverskey (1979)提出展望理論及價值函數，說明投資人在不確定性的情形下，從事各種投資決策，會有不理性的決策行為產生。為消弭此一不理性的決策行為，投資人嘗試藉由程式交易軟體，來降低投資人的不理性決策干預。本文的主要研究目的乃採用股票技術指標作為投資人心理情緒之間接代理變數，試圖以非線性的STAR模型求出投資人心理情緒的多空轉折點，作為程式交易參數之設定標準，以科學化的方式建構報酬最佳化交易模型。實證結果發現：經樣本內回溯測試及樣本外預測結果均顯示：若在心理線(PSY)的門檻值大於90.15時執行賣出交易與在相對強弱指標(RSI)的門檻值小於7.53時執行買進交易，則獲得正報酬比率均為100%，且門檻值買進條件RSI<7.53在空頭時期更具篩選效果。此結果顯示：部分的股票技術分析指標可以有效的詮釋投資人心理情緒的多空轉折點。 Kahneman and Tverskey (1979) had supposed the prospect theory and value function to describe the human decision rule in an uncertain situation and explain the decision due to the investor irrationality. For excluding the irrational decision, the financial engineering is trying to reduce intervention of human irrationality by the program-trading software. In view of these points, this paper use technical indicators employed as instrument variables of the investor sentiment. The purpose of this study is an attempt to find out the turning point between long and short investment in the investor sentiment by smooth transition autoregressive (STAR) model and these threshold values are smooth transition autoregressive (STAR) model and these threshold values are employed in trading function. By back-testing in the sample and forecasting out of the sample, the empirical results revealed that both the ratio of positive return were 100%, while psychology line (PSY) is over 90.15 and relative strength index (RSI) is under 7.53, besides, the screening effect of the latter is better in the bull period. The results imply the partial stock technical analysis can effectively explain the turning point between long and short investment in the investor sentiment.
