整合多面向標準評比與天際線篩選方法於共同基金之分析及預測─以投資國內股票型共同基金為例

投資人常需要蒐集並分析相關的財經資訊，以增加投資獲利的機會。隨著網際網路的蓬勃發展，投資人能輕易取得大量的基金相關資料，隨之而來的資訊超載也日益嚴重。因此，有系統地建立一套機制來篩選並分析持續累積的財經資料，一直是有趣但具挑戰性的研究課題。本論文以國內『中華民國證券投資信託暨顧問商業同業公會』網頁上每月所公佈的基金績效評比表為主要資料來源，建立一套共同基金之分析及預測系統，以作為投資人買賣國內股票型基金時的輔助工具。系統分為兩部分，首先以「強勢基金篩選系統」挑選各期表現優異的基金，包括以「四四三三法則」找出投資報酬率名列前茅的基金，並利用「天際線支配法則」找出風險考量下績效評比具優勢的基金。接著，「基金分析與趨勢預測系統」依使用者指定的時段，擷取相關各期強勢基金的資訊，以分析個別基金的綜合績效與未來持續性，或評估強勢基金的整體表現，以預測未來市場的多空趨勢。實驗結果顯示，在所篩選基金的後續表現上，我們的方法優於一般常用的篩選法則，趨勢預測也有近60%的準確率；而天際線篩選的加速方法可提昇效率達20%以上。

關鍵字

資料探勘；天際線；強勢基金；四四三三法則；共同基金

並列摘要

Investors often need to collect and analyze financial and economic reports for taking the opportunity of making money. As the booming development of Internet, it is easy for investors to get the related data, but the information overloads is getting worse. Therefore, a systematic way to build a mechanism for the filtering and analysis of such data has become an interesting but challenging topic. This thesis takes the monthly reports of “Fund Performance Evaluation Form” announced by “Securities Investment Trust Consulting Association of the R.O.C.” (SITCA) as the main resource to build a system for the analysis and prediction of mutual funds and help investors do business. Our system consists of two parts. The first subsystem finds out the strong funds that perform well in two phases. One is based on “4433 Rules” to pick up the funds with the best performance in the gain rate, while another applies the skyline concept to choose the funds without being dominated by any others. The second subsystem then retrieves the information of the strong funds within the user-specified time period to analyze their performance and durability in the near future. Integrating the analytical results of all the strong funds further enables the prediction of the entire marketing trend in the near future. The experiment results show that our method outperforms the ordinary rules in the subsequent performance of the filtered results. The prediction also achieves almost 60% in accuracy. Moreover, our method for skyline computation speeds up above 20%.

並列關鍵字

Data Mining ； Strong Funds ； Skyline ； 4433 Rules ； Mutual funds

參考文獻

[17] 郭維裕, 李愷莉, “台灣共同基金短期績效持續性的研究─以「漂移者─停駐者」模型為例,” 中央研究院經濟研究所經濟論文, Vol.34(4):469-504, 2006.

[1] G. Adomavicius and Y.O. Kwon, “New Recommendation Techniques for Multicriteria Rating Systems, ” IEEE Intelligent Systems, Vol.22:48-55, MAY/JUNE 2007.

[2] S. Borzsonyi, D. Kossmann and K. Stocker, “The Skyline Operator,” in Proc. of the 17th International Conference on Data Engineering, 2001.

[5] G. Linden, B. Smith and J. York, “Amazon.com Recommendations Item-to-Item Collaborative Filtering,” IEEE Internet Computing, Vo.7(1):7680 , January. 2003.

[6] X. Lin, Y. Yuan, W. Wangnicta and S. Wales, “Stabbing the Sky : Efficient Skyline Computation over Sliding Windows,” The 21st International Conference on Data Engineering, 2005.

國際替代計量

整合多面向標準評比與天際線篩選方法於共同基金之分析及預測─以投資國內股票型共同基金為例

全文下載

主題瀏覽