多市場多商品程式交易績效之決策支援平台的整合設計與實作

The Integrated Design and Implementation of Decision Support Platform of Multi-Market and Multi-Stock Program Trading Performance

楊宗慶(Zong-cing Yang)

指導教授 ： 許智誠

國立中央大學/管理學院/資訊管理學系/碩士(2015年)

摘要

股票操作屬於高風險高報酬的投資行為，為降低股票投資風險與追求穩定的投資報酬，股票市場上存在許多的股票分析方式 (基本、技術、產業、趨勢等分析方式)，衍伸出運用不同的分析組合以及提供有價值決策的交易策略。

大部分的投資者認為實際交易是一件很困難的事情，因為投資者可能被許多因素影響投資決策 (黃培琳, 2011)，因此許多投資者紛紛將交易轉介於程式上進行程式交易。程式交易是以明確的程式語言定義買賣條件由程式自動執行交易，程式交易可透過歷史的股價進行回測調整，也可以避免人為投資所造成的問題。

程式交易策略近年來已逐漸商業化，且目前投資者普遍將程式交易策略的績效記錄在紙本、Excel、Word 或整理在網路部落格 (周志強, 2014)，缺乏統一的架構與平台。投資者面對這些各式各樣的程式交易策略，難以做出績效的分析與比較，對於程式交易策略適用特性 (市場、股票類型、期間等) 也難以捉摸，導致投資者對程式交易策略的穩定性抱持著懷疑的態度，進而難以降低投資風險與追求更高的投資報酬。

本研究透過銜接與整合現有軟體，建立自動化程式交易績效管理工作流程，確保程式交易策略的績效紀錄可以穩定產出並妥善儲存，藉由產出的績效紀錄建立資料倉儲，最後再以資料倉儲建置程式交易決策支援平台 (myPTA)，提供投資者程式交易策略的績效分析與比較，支援投資者做出投資決策、降低股票投資風險、追求更穩定的報酬。

藉由程式交易績效管理與決策支援平台的驗證，確認投資者確實可以藉由本研究所提的自動化程式交易績效管理工作流程，有效管理程式交易的績效。並透過本研究提出的統一管道與平台，進行多市場多商品多策略的比較與分析、程式交易策略的分享，藉由社群的力量進行協同合作。

關鍵字

系統整合規劃 ； 程式交易 ； 資料倉儲 ； 工作流程

並列摘要

Stock operations are high risk, high reward investment behavior to reduce stock investment risk and seeking stable return on investment, there are many way to analyze stocks on the stock market. Derived trading strategies that using different combinations of analysis and providing valuable trading decisions.

Most investors believe that the actual transaction is a very difficult thing, because there are many factors which may influence the investment decision. Therefore, so many investors are turn to using programing trading. Programing trading use Program languages to define condition of buy (sell) and trade automatically, program trading can be tested through historical price adjustment, and also avoid problems caused by human investment.

Program trading strategies has been commercialized, and now people will record the performance of program trading strategies on paper, Excel, Word or blog, but these records are lacking a unified architecture and platform. It is difficult to make analysis and comparison of performance when Investors face a wide variety of program trading strategies. This will lead investors to doubt the stability of program trading strategies, and hard to reduce investment risk and higher return on investment.

This research established an automation program trading performance management work flow through convergence and integration existing software, ensure program trading strategies of performance records can stable output and properly store. We use performance records to establish the data warehouse, and finally to build a program trading decision support platform by this data warehouse (myPTA) that can provide performance analysis and comparison of program trading strategies to support investors to make investment decisions, reduce stock investment risk and the pursuit of more stable pay.

Through verifying performance management and decision support platform to ensure investors can use automation program trading performance management work flow which we established in our research to manage program trading performance effectively. And through the proposed unified pipeline and Platform for analysis and comparison of multiple market, multiple products and multiple strategies, sharing of program trading strategies and collaboration by the strength of community.

並列關鍵字

Enterprise Application Integration ； Program Trading ； Data Warehouse ； Workflow

參考文獻

邱博政. (2008). 基於工作流程與通用隨插即用技術下之整合性軟體開發架構. 國立交通大學資訊科學與工程研究所碩士論文.

Christopher C. Shilakes, Julie Tylman. (1998). Enterprise Information Portals. Merrill Lynch.

Coalition, W. M. (1999). Terminology & Glossary. The Workflow Management Coalition.

Murphy, J. J. (1999). Technical Analysis of the Financial Markets. New York Institute of Finance.

PardoRobert. (2012). The Evaluation and Optimization of Trading Strategies. John Wiley & Sons, Ltd.