本研究利用十四種技術指標,將技術指標計量化建構複合指標模型,在模型中加入總體經濟指標控制變數提升對於模型整體對於應變數的解釋能力,研究探討各種技術指標對於台股加權指數的影響能力,並且檢驗單一技術指標對於加權指數的預測能力。以上都成立才對於每一種技術指標搭配一種複合指標,檢驗兩種指標的複合搭配對於台股加權指數的顯著程度,最後以計量模型預測對於加權指數報酬率的準確度與誤差,希望能給投資人指標搭配上的建議。 研究結果顯示單一技術指標對於加權指數報酬率雖有顯著影響能力,但是不一定有良好預測能力。十四種技術指標中有八種具有預測能力、六種不具有預測能力;利用本研究方法找出四十四組複合指標都具有顯著影響能力,其中有十八組具有預測能力,十四組不僅具有預測能力且比單一技術指標還要優良。在長、短區間迴歸做樣本外預測以積量指標與乖離率的複合指標模型最好,準確度最高為100%
In this study use fourteen kinds of technical indicators to construct combined indicator forecast model, to better predicting index premium in each models put macroeconomic variables to improve overall explain ability. This study investigate relationships between technical indicators and stock index, and test technical indicators whether do predict stock index effectively. Moreover, we combine two indicators to enhancing the relation of indicators and stock index, and test whether combined models do predict better than single indicator models. The empirical study shows each indicator do effect stock index significantly, but predict. Hence that, combine two indicators to find out which sets are efficiently. The findings of study is cumulative volume indicator combine bias indicator is the best performance model, which accuracy rate and errors are excellent in all models.