台灣股市之技術指標分析與量關係之實証研究

Abstract

由於國內近年經濟成長迅速，且為求進軍世界金融中心，故對於一般金融 投資分析需求更加殷切。 有鑑於國內投資方面以股市居多， 故利用現今 電腦強大之計算能力，對於股市之分析方法做一實証性研究。於本論文中 ， 本研究利用電腦對於一般常用之股市分析之技術指標，以統計分析之 方式驗証。 其資料來源為最近五年內之証券交易資料，經過轉換之後進 行分析研究。 經過多次的程式驗証，本研究得到下面的成果:1. 若利用 成交量代入原先以成交價計算之技術指標， 有時可以得到較原先僅利用 成交價計算之指標更加準確的結果。 2. 技術指標係隨『指標參數』、『 時間』、『買入 / 賣出』之三維動態變化過程； 買入及賣出應選用不同 參數之技術指標，且在不同時期亦應改變其指標參數，方可獲致最佳分析 效益。3. 利用一組均價所產生之 Bias 做為分析時， 其均價間之 Variance 在 Local Max/Min 附近時，經常為股價長期變動點之起始點 。4. 累積成交量所產生之壓力 / 支撐效應並未如想像中明確。於論文之 最後，本研究提出「電腦輔助股市分析模型」。 由於技術指標在不同時 期應利用不同之參數做為買 / 賣之參考，故利用電腦之強大運算能力， 每日即時針對所有可資利用之技術指標及不同參數， 以統計方法將其歷 史準確率做為評判準則，而獲致一快速分析各不同股票之方法， 俾使分 析人員能在有限之時間內，快速掌握各不同股票之間之變化，以利分析決 策之用。 Due to the economical growth in Taiwan and the ambition to become Asia Financial Center, the need of commercial analysis technique is more important than before. This research uses computer-aided analysis method to get an empirical study of Technical Index Analysis with Price-Volume Relationship in Taiwan Stock Market. In this research, five years of Taiwan stock market data was taken to build an empirical study and concluded the following results. First, technical index may be represented by exchange stock volume instead of the stock price, and which may sometimes derive better accuracy of the output results. Second, the technical index analysis may be influenced by the index parameter, i.e. different parameter may reach different results and accuracy. Third, for each technical index the accuracy may vary dramatically du ring different periods of time, i,e. no technical index may be effective forever. Forth, the buy-signal accuracy and sell-signal of one technical index are different and the analysts should take different parameters while making buying and selling decisions. Fifth, a study on advanced bias analysis shows that the long-term variation points may match the local min/max points of moving average price sets' standard deviation. Sixth, pressure/support analysis derived from accumulated stock volume shows no tight ly coupled relationships between long-term variation points and accumulated volume. With the use of the research results, the end of this thesis shows a 「 Computer Aided Stock Analysis System Model (CASAS) 」. The analysts may take the advantage of the computing ability of CASAS to take all the variation of Technical Index into consideration during analysis processes.