透過您的圖書館登入
IP:18.222.111.211
  • 學位論文

應用平行K-means演算法建構股市決策支援系統於Hadoop平台

Using Parallel K-means Clustering to Construct the Stock Decision Support System Based on Hadoop Platform

指導教授 : 李國誠 李彥賢

摘要


股市投資一直以來都是一項熱門的投資方式,其投資方法主要分成基本分析、技術分析、籌碼分析。在技術分析中,投資者會根據不同的技術指標來投資股票,但技術指標種類繁多,時常讓投資者不知道該使用哪種技術指標來投資,導致投資結果並不理想。本研究以TEJ資料庫做為股市資料來源,將以Hadoop平台為基礎讓股票先透過技術指標公式平行運算後,接著將K-means演算法套用於MapReduce框架上,藉此將股票作分群並同時提高運算效率,最後將分群結果定義決策後,再推薦投資者作買進或賣出之決策。本研究之決策支援系統將採用技術指標的公式及K-means演算法來建構其中的模式管理子系統部分,並將分析出的結果呈現出來供投資者作參考。

並列摘要


Stock market investment has always been a popular way to invest. The approach is divided into fundamental analysis, technical analysis, and chip analysis. In technical analysis, investors will invest in stocks based on different technical indicators, but there are various kinds of technical indicators that usually make the investors be confused of using which technical indicators to invest, leading to investment results are not satisfactory. In this study, the stock market data is sourced from TEJ database. This paper will be based on Hadoop platform, let the stock after the first through the parallel computing technology indicators formula, and set the K-means algorithm into the MapReduce framework, to cluster stocks and improve operation efficiency at the same time. At last, define the clustering results, and then recommend the investors to make decisions for buying or selling. The decision support system of this paper will be used technical indicators formulas and K-means algorithm to construct a section in which model management system, and analyze the results for investors to reference.

參考文獻


趙晟鈞. (2012). 以約略集合與灰關聯結合技術分析建置投資組合策略-以台灣上市公司為例. 嶺東科技大學.
高榮泰. (2009). 使用具代表性之特徵預測股市趨勢. 國立台灣大學.
李顯儀, & 吳幸姬. (2009). 技術分析資訊對共同基金從眾行為的影響. 臺大管理論叢, 20, 227–260.
陳金泉. (2010). 應用資料探勘找出金融海嘯下台灣股市的避險操作方式. 大同大學.
簡玠忠. (2013). 基於 Hadoop 框架建立巨量資料分析處理模型研究.

被引用紀錄


黃筠珺(2018)。流動性與VIX指數對股市動能策略市場之研究〔碩士論文,中原大學〕。華藝線上圖書館。https://doi.org/10.6840/cycu201800058

延伸閱讀