臺灣博碩士論文加值系統

大數據分析在各行各業間的應用漸趨廣泛，本論文探討應用數據分析能否改善基金績效。本文利用群體智慧的案例式推理系統調整投資組合，並檢視基金投資組合透過推理系統改變後的基金績效與原始的基金績效是否有顯著差異。
案例式推理系統(Case-Based Reasoning system)是一種推薦系統，透過尋找相似案例，並利用其解決問題的方法來試圖解決新的問題。本文利用兩種找尋相似案例的推理法進行投資組合的改善，第一種是以數學分析的方式選取鄰近組合，第二種是以人工的方式選取鄰近組合，接下來依據推理系統應用的五個步驟：檢索、重用、修訂、審查和保留。首先我們在檢索和重用步驟中選出鄰近組合，接著我們利用修訂步驟來過濾我們的鄰近組合，最後在審查與保留步驟中，將所選鄰近組合的投資組合簡單平均，建立出新的投資組合，並檢測這兩種選擇鄰近組合的方法所調整的投資組合，對於基金績效是否有顯著差異。研究結果顯示，利用這兩種群體智慧的方法進行基金持股的選擇，能使基金投資組合多元化，同時使基金績效顯著提升。

The Big data analysis has been used more extensively in various industries. This paper explored whether the big data analysis can help fund managers to improve fund performances. This study used the Case-Based Reasoning system of the collective intelligence to develop new investment strategies. The thesis examined if there was a significant difference between the fund performance with original investment strategies and the one with improved strategies in which share holdings are changed by the Case-Based Reasoning system.
The Case-Based Reasoning system is a recommendation system that tries to solve new problems by looking for solutions in similar cases. Therefore, this study used two methods to find similar cases, namely mathematical and artificial methods. According to the five steps of Case-Based Reasoning system: the retrieve, reuse, revise, review and retain steps, this study firstly singled out the neighborhoods through the first two steps- retrieval and reuse; then picked out the relatively diversified neighborhoods with better performances in the revise step; finally, hence built a new portfolio by averaging out the portfolios of selected neighborhoods in the last two steps- review and retain. This paper tested whether there were different fund performances between these two methods. The thesis provided evidence that the usage of the two fund-based collective intelligence methods can develop more diversified portfolio and significantly improve fund performances.

目錄
摘要 I
Abstract II
圖目錄 V
表目錄 VI
第一章、緒論 1
第一節、研究動機 1
第二節、研究目的 2
第三節、研究架構 3
第二章、文獻探討 5
第一節、CBR案例式推理 5
第二節、假說 8
第三章、研究方法及樣本 9
第一節、研究方法 9
第二節、案例篩選系統原理 10
第三節、模型 12
第四節、樣本資料介紹 13
第五節、樣本資料處理 13
第四章、模型實證與資料分析 15
第一節、敘述性統計 15
第二節、研究實證(檢索) 15
第三節、研究實證(重用) 17
第四節、研究實證(修訂) 17
第五節、研究實證(審查與保留) 18
第六節、研究實證(整體) 19
第七節、研究實證(迴歸模型) 21
第五章、實證與結果 23
參考文獻 24
附錄 25

1. Musto, C., Semeraro, G., Lops, P., de Gemmis, M.
and Lekkas, G. (2015). Personalized finance advisory
through case-based recommender systems and
diversification strategies. Decision Support Systems.
77, 100-111.
2. Smyth, B., McClave, P. (2001). Similarity vs. Diversity,
Case-Based Reasoning Research and Development. Springer.
347-361
3. Trewin, S. (2000). Knowledge-based recommender systems.
Encyclopedia of library and information science.
69(Supplement 32), 180.
4. Montaner, M., López, B. & de la Rosa, J. (2002).
Improving case representation and case base maintenance
in recommender agents. Advances in Case-Based Reasoning
.1-17.
5. Ruslan, Goyenko. 2013). Mutual Fund's R2 as Predictor of
Performance. Review of Financial Studies. 26, 667-694