資訊管理學報
陳振東;謝政翰;
頁: 153-177
日期: 2019/04
摘要: 近年來，利用智慧數據分析方法以預測股價乃是金融科技（Financial Technology; FinTech）領域的重要議題。然而，有許多的技術指標以及人為主觀因素會影響股價的漲跌預測，因此必須有效掌握重要的影響指標，才能提高股價漲跌預測的正確率。為此，本研究透過技術指標的篩選程序，使用四種機器學習演算法進行股價漲跌的預測分析，進而篩選重要的技術指標。此外，由於技術指標的屬性及人為的主觀判斷具有不確定性與模糊性，因此本研究應用模糊推論方法建構模糊推論系統以進行股價漲跌的預測，並提出股價漲跌幅度區間的預測方法。最後，本研究針對三家公司的股價資料進行實證分析，研究結果顯示股價漲跌預測的正確率都達82.13%以上，三家公司股票價格的漲跌幅度區間涵蓋真實股價的平均預測正確率都高達83%以上。由此可知，本研究所提出的模糊推論預測系統不僅具有學理基礎，同時能夠有效預測股票的漲跌趨勢及漲跌幅區間，對投資人具有實務應用的價值與貢獻。
關鍵字: 金融科技;股價漲跌預測;機器學習演算法;模糊推論預測系統;
Abstract: Purpose-The purpose of this study is to propose a fuzzy inference forecasting system to predict the variation of stock price of each company and the stock price fluctuation range. Design/methodology/approach-In this study, four machine learning algorithms are used to predict the stock prices and select the important technical indicators. And then, this study applies fuzzy inference to construct a fuzzy inference system to predict stock price fluctuation based on the critical technical indices. Findings-The results of case study showed that the accuracy of the stock price fluctuation is more than 82.13% for three companies. In addition, this study also proposes a new method for predicting the stock price fluctuation range. The research results show that the prediction accuracy rate of stock price intervals are more than 83% for three companies. Research limitations/implications-This study focused on the numeral data of technical indicators but not non-numeral data in the fuzzy inference system to predict stock price fluctuation. Future research can combine different data types to construct the prediction model of stock price. Practical implications - According to the fuzzy inference forecasting system, investors can use only the five technical indices to predict the fluctuation trend and the interval of the stock price. The prediction accuracy rate of stock price intervals are good enough for investors. Originality/value-The fuzzy inference forecasting system proposed in this study not only have the academic values, but also can effectively predict the fluctuation trend and the interval of the stock price, and the contributions of practical applications for investors.
Keywords: FinTech;stock price forecasting;machine learning;fuzzy inference forecasting system;
瀏覽次數: 20761 下載次數: 984
引用 導入Endnote
頁: 153-177
日期: 2019/04
摘要: 近年來，利用智慧數據分析方法以預測股價乃是金融科技（Financial Technology; FinTech）領域的重要議題。然而，有許多的技術指標以及人為主觀因素會影響股價的漲跌預測，因此必須有效掌握重要的影響指標，才能提高股價漲跌預測的正確率。為此，本研究透過技術指標的篩選程序，使用四種機器學習演算法進行股價漲跌的預測分析，進而篩選重要的技術指標。此外，由於技術指標的屬性及人為的主觀判斷具有不確定性與模糊性，因此本研究應用模糊推論方法建構模糊推論系統以進行股價漲跌的預測，並提出股價漲跌幅度區間的預測方法。最後，本研究針對三家公司的股價資料進行實證分析，研究結果顯示股價漲跌預測的正確率都達82.13%以上，三家公司股票價格的漲跌幅度區間涵蓋真實股價的平均預測正確率都高達83%以上。由此可知，本研究所提出的模糊推論預測系統不僅具有學理基礎，同時能夠有效預測股票的漲跌趨勢及漲跌幅區間，對投資人具有實務應用的價值與貢獻。
關鍵字: 金融科技;股價漲跌預測;機器學習演算法;模糊推論預測系統;
A Study on Application of Machine Learning and Fuzzy Inference for the Prediction of Stock Price
Abstract: Purpose-The purpose of this study is to propose a fuzzy inference forecasting system to predict the variation of stock price of each company and the stock price fluctuation range. Design/methodology/approach-In this study, four machine learning algorithms are used to predict the stock prices and select the important technical indicators. And then, this study applies fuzzy inference to construct a fuzzy inference system to predict stock price fluctuation based on the critical technical indices. Findings-The results of case study showed that the accuracy of the stock price fluctuation is more than 82.13% for three companies. In addition, this study also proposes a new method for predicting the stock price fluctuation range. The research results show that the prediction accuracy rate of stock price intervals are more than 83% for three companies. Research limitations/implications-This study focused on the numeral data of technical indicators but not non-numeral data in the fuzzy inference system to predict stock price fluctuation. Future research can combine different data types to construct the prediction model of stock price. Practical implications - According to the fuzzy inference forecasting system, investors can use only the five technical indices to predict the fluctuation trend and the interval of the stock price. The prediction accuracy rate of stock price intervals are good enough for investors. Originality/value-The fuzzy inference forecasting system proposed in this study not only have the academic values, but also can effectively predict the fluctuation trend and the interval of the stock price, and the contributions of practical applications for investors.
Keywords: FinTech;stock price forecasting;machine learning;fuzzy inference forecasting system;
瀏覽次數: 20761 下載次數: 984
引用 導入Endnote