臺灣博碩士論文加值系統

基金投資一直是國內投資市場中相當受歡迎的一種投資商品，然
而近年來受制於金融海嘯的衝擊之下，全球基金淨值大幅下滑，許多
國內投資人於基金投資中遭受重大的損失，而投資人遭受損失後，退
場觀望，所幸各國政府的財經政策引導下，景氣逐漸恢復，投資者對
基金商品又恢復信心，而如何選擇穩健的基金標的又成為市場探討的
話題。
本文的主旨在於探討國內的股票型基金在2009 年的盈虧狀況，並
分析影響基金盈虧的因素，文中計算每支基金在不同區段內的績效得
分並找出績效良好的基金，依照績效得分的排名，將基金分群，並使
用回歸分析法等資料探勘技術，探討基金盈虧與投資標的，資產配置，
基金基本狀況之間的關係。最後透過問卷來確認論文的研究流程，期
望透過上述手法，能提供投資人日後選擇基金有效的建議。

Mutual fund is one of the popular investment tools in domestic financial market. Recently, many investors had suffered a great loss after financial tsunami. Many investors reduced their asset to invest mutual funds. However, with the help from the government policy, the global financial market has recovered in 2009. Mutual fund investment became a popular investment channel again. The main point of this thesis is to explore the performance of mutual funds in 2009 and to analyze the factors that affect the performance of mutual funds. This thesis helps us to find the outstanding mutual funds by calculating the net asset value (NAV) return between different time intervals and then group the mutual funds by ranking. Furthermore, it applied data mining technology, such as regression analysis, to discuss the relationship between the performance of mutual funds, asset allocation, and investment target. The last process of this thesis is to confirm the research process by questionnaire survey. We hope this research can provide suggestions and reference rules to investors.

摘要 i
Abstract iii
LIST OF CONTENTS v
LIST OF TABLES ix
LIST OF FIGURES xi
Chapter 1 Introduction 1
1.1 Background 1
1.1.1 The domestic status of mutual funds 2
1.1.2 The technology approach: data mining 3
1.2 Research problem 4
1.3 Research motivation 5
1.4 Research purpose 5
1.5 Research scope 6
1.6 Research limitation 6
1.7 Research contribution 7
1.8 Thesis outline 7
Chapter 2 Literature review 9
2.1 Modern portfolio theory 9
2.2 Fund performance indicators related literature 10
2.2.1 Net asset value return 10
2.2.2 Beta coefficient 10
2.3 Recently mutual fund performance literature 11
2.4 Data mining technology 17
2.4.1 Association rules 17
2.4.2 Classification 18
2.4.3 Cluster analysis 19
2.4.4 Statistic analysis 20
2.4.4.1 Regression analysis 20
2.4.4.2 Factor analysis 21
2.4.4.3 Correlation analysis 22
2.5 Cross-industry standard process of data mining 22
2.5.1 The phase of cross-industry standard process of data mining 23
2.6 Recently literature of data mining technology 27
Chapter 3 Methodology 31
3.1 Research process 32
3.2 Research model 33
3.3 Research method 37
3.3.1 Data collection 38
3.3.2 Data description 38
3.3.3 Data exploration 39
3.4 Data preparation 40
3.4.1 Clean data 41
3.4.2 Select data 41
3.4.3 Construct data 41
3.4.4 Transform data 47
3.5 Modeling 47
3.5.1 Model software select 47
3.5.2 Model technology select 48
3.5.2.1 SPSS clementine 12.0 parameters setting 48
3.5.3 Model test design 52
3.6 Evaluate by questionnaire survey 54
3.6.1 Questionnaire administration mode 55
3.6.2 Questionnaire indicators 57
3.6.2.1 The reliability 57
3.6.2.2 The validity 58
3.6.3 Questionnaire survey method 60
Chapter 4 Research finding 61
4.1 Data mining research finding 61
4.1.1 The Taiwan small/mid-cap equity mutual funds 61
4.1.1.1 The research finding of Taiwan small/mid-cap
equity mutual funds (node B1) 64
4.1.1.2 The research finding of Taiwan small/mid-cap equity mutual funds (node C1) 65
4.1.1.3 The research finding of Taiwan small/mid-cap
equity mutual funds (node C2) 66
4.1.2 The Taiwan large-cap equity mutual funds 71
4.1.2.1 The research finding of Taiwan large-cap equity mutual funds (node B2) 73
4.2 Questionnaire research finding 78
4.2.1 Investment background and experience 79
4.2.2 Research finding of data mining fit model 84
4.3 Validity analysis 93
4.4 Reliability analysis 95
4.5 Path analysis 96
4.6 Verification analysis of hypotheses 98
4.7 Finding and discussion 100
Chapter 5 Conclusion 109
Chapter 6 Managerial implication 115
Reference 119
Appendices A: Tables 123
Appendices B: Figures 149
Appendices C: Questionnaires 157

LIST OF TABLES
Table 3.1 Research scope 31
Table 3.2 Variables 40
Table 4.1 Score and classified table of Taiwan small/mid -cap equity mutual funds (good funds) 62
Table 4.2 Score and classified table of Taiwan small/mid -cap equity mutual funds (worse funds) 63
Table 4.3 Model summary of node B1 64
Table 4.4 Model summary of node C1 66
Table 4.5 Descriptive statistics of node C2 67
Table 4.6 Pearson correlations of node C2 68
Table 4.7 Model summary of node C2 69
Table 4.8 ANOVA of node C2. 70
Table 4.9 Coefficients statistics of node C2 71
Table 4.10 Score and classified table of Taiwan large
-cap equity mutual funds 72
Table 4.11 Descriptive statistics of node B2 74
Table 4.12 Pearson correlations of node B2 75
Table 4.13 Model summary of node B2 76
Table.4.14 ANOVA of node B2 77
Table 4.15 Coefficients statistics of node B2 78
Table 4.16 Result of investment background and experience of investors 80
Table 4.17 Descriptive statistics of questionnaires (questions) 85
Table 4.18 Descriptive statistics of questionnaires (scopes) 86
Table 4.19 Pearson correlations of model I. 87
Table 4.20 Pearson correlations of model II 88
Table 4.21 Model summary of model I 89
Table 4.22 Model summary of model II 89
Table 4.23 ANOVA of model I 90
Table 4.24 ANOVA of model II 91
Table 4.25 Coefficients statistics of model I 92
Table 4.26 Coefficients statistics of model II 92
Table 4.27 KMO and Bartlett's test (model I) 93
Table 4.28 KMO and Bartlett's test (model II) 94
Table 4.29 Component matrix of questionnaires 94
Table 4.30 Cronbach’s α of model scope 96
Table 4.31 Path analysis table 97
Table 5.1 Verification result 111


LIST OF FIGURES
Figure 2.1 The phase of CRISP-DM model 24
Figure 3.1 Research process 31
Figure 3.2 Data mining fit model 33
Figure 3.3 Research data process 37
Figure 3.4 The binary tree 43
Figure 3.5 Fund rank list tree 46
Figure 3.6 Setup the independent variables and the dependent variable 49
Figure.3.7 Setup method for linear regression 50
Figure 4.1 The path analysis of data mining fit model 97

[1] Securities Investment Trust & Consulting Association, http://www.sitca.org.tw/.

[2] M.-S. Chen, J. Han, and P. S. Yu, “Data Mining：An Overview from a Database Perspective,” IEEE Transactions on Knowledge and Data Engineering, Vol.8, No. 6, pp. 866-883, 1996.

[3] Pieter Adriaans and Dolf Zantinge, “Data Mining,” Addison Wesley, New York, 1996.

[4] Jeffery W. Seifert, “Data Mining: An Overview,” CRS Report for Congress, Received through the CRS Web, December 12, 2004.

[5] Hui, SC; Jha, G, “Application of data mining techniques for improving customer services,” International journal of computer applications in technology, Vol.14, No.1-3, pp.64-77, 2001.

[6] Kantardzic, Mehmed, “Data Mining: Concepts, Models, Methods, and Algorithms,” John Wiley & Sons, 2003.

[7] Markowitz, Harry M, “Portfolio Selection,” Journal of Finance, Vol.7, No.1, pp.77-91, 1952.

[8] Edwin J. Elton and Martin J. Gruber, "MPT, 1950 to date," Journal of Banking & Finance 21, pp.1743-1759, 1997.

[9] CFA Institute, “Derivatives and Alternative Investment, ” Boston: Pearson Custom Publishing, 2008.

[10]Levinson, Mark. “Guide to Financial Markets,” London: The Economist, pp.145–146, 2006.

[11]Kosowski, R; Timmermann, A; Wermers, R; White, H, “Can mutual fund stars really pick stocks? New evidence from a bootstrap analysis,” Journal of Finance, Vol.61, No.6, pp.2551-2595, 2006.

[12]Frazzini, A; Lamont, OA , “Dumb money: Mutual fund flows and the cross-section of stock returns,” Journal of Financial Economics, Vol.88, I.2, pp.299- 322, 2008.

[13]Ferson, W; Henry, TR; Kisgen, DJ ,“Evaluating government bond fund performance with stochastic discount factors,” Review of Financial Studies , Vol.19, No.2 , pp.423-455, 2006.

[14]Avramov, D; Wermers, R, “Investing in mutual funds when returns are predictable,” Journal of Financial Economics, Vol.81, No.2, pp.339-377, 2006.

[15]Friesen, GC; Sapp, TRA, “Mutual fund flows and investor returns: An empirical examination of fund investor timing ability,” Journal of Banking and Finance, Vol.31, pp.2796-2816, 2007.

[16]Busse, JA; Irvine, PJ, “Bayesian alphas and mutual fund persistence,” Journal of Finance, Vol.61, No.5, pp.2251-2288, 2006.

[17]James, C; Karceski, J, “Investor monitoring and differences in mutual fund performance,”Journal of Banking and Finance,Vol.30,No.10, pp.787-2808, 2006.

[18]Cuthbertson, K; Nitzsche, D; O'Sullivan, N, “UK mutual fund performance: Skill or luck,” JOURNAL OF EMPIRICAL FINANCE, Vol.15, No.4, pp.613-634, 2008.

[19]Basak, S; Pavlova, A; Shapiro, A,“Optimal asset allocation and risk shifting in money management,” Review of Financial Studies, Vol.20, p.p.1583-1621, 2007.

[20]Ramamohanarao, K; Fan, HJ, “Patterns based classifiers,” World Wide Web - internet and web information systems, Vol.10, No.1, pp.71-83, 2007.

[21]R. Agrawal; T. Imielinski; A. Swami: “Mining Association Rules between Sets of Items in Large Databases,” SIGMOD Conference, pp.207-216, 1993.

[22]Jiawei Han, Micheline Kamber, “Data Minging: Concepts and Techniques,” Morgan Kaufmann Publishers, 2001.

[23]Anderberg, Michael R.,“Cluster Analysis for Applications,”Academic Press, New York, 1973.

[24]Sykes, A. O.,“An Introduction to Regression Analysis,”Chicago Working Papers (September), No.20, 1993.

[25]Velicer, W. F. & Jackson, D. N., "Component analysis versus common factor analysis: Some issues in selecting an appropriate procedure," Multivariate Behavioral Research, Vol.25, No.1, pp.1-28, 1990.

[26]Colin Shearer, “The CRISP-DM Model: The New Blueprint for data Mining,” Journal of Data Warehousing, Vol.5, No.4, pp13-22, 2000.

[27]Pete Chapman (NCR), Julian Clinton (SPSS), Randy Kerber (NCR), Thomas Khabaza (SPSS), Thomas Reinartz (DaimlerChrysler),Colin Shearer (SPSS) and Rüdiger Wirth (DaimlerChrysler), “CRISP-DM 1.0 Step-by-step data mining guide,” The CRISP-DM consortium, 2000.

[28]The CRISP-DM consortium, http://www.crisp dm.org/SIG/ index.htm.

[29]Liao, SH; Ho, HH; Lin, HW, “Mining stock category association and cluster on Taiwan stock market,” Expert systems with applications, Vol.35, No.1-2, pp.19-29, 2008.

[30]Hayward, S, “Genetically optimized artificial neural network for financial time series data mining,” simulated evolution and learning, proceedings, Vol.4247, pp. 703-717, New York, 2006.

[31]Savvides, A; Promponas, VJ; Fokianos, K, “Clustering of biological time series by cepstral coefficients based distances,Pattern recognition,” Vol.41, No.7, pp.2398-2412, 2008.

[32]Ellis Horowitz, Sartaj Sahni, Dinesh Mehta, “fundamentals of data structures in C++,” W. H. Freeman and Company, 1995.

[33]Cronbach, L. J., “Coefficient alpha and the internal structure of tests,” Psychometrika, Vol.16, No.3, pp.297-334, 1951.

[34]Maria Teresa Siniscalco and Nadia Auriat“Questionnaire design,” Quantitative research methods in educational planning , Moudle.8, 2005.

[35]Cook, T., & Campbell, D. T., Quasi-experimentation: Design and analysis issues for field settings,” Boston: Houghton-Mifflin, 1979.

[36]Lee J. Cronbach and Paul E. Meehl, “Construct Validity in Psychological Tests,” Psychological Bulletin, Vol.52, pp.281-302, 1955.

[37]Likert, R., “A Technique for the Measurement of Attitudes,” Archives of Psychology, Vol. 140, pp.1-55, 1932.

[38]H.Kaiser, “An index of factorial simplicity,” Psychometrika, Vol.39, pp.31-36, 1974.

[39]R. F. Devellis, “Scale development : theory and application,” Newbury Park, CA: Sage,1991.

[40]Goetzmann, William N and Nadav Peles, “Cognitive Dissonance and mutual fund investors,” Journal of Financial Research, Vol.20, pp.145-158, 1997.

[41]S. Chaudhuri and U. Dayal, “An overview of data warehouse and OLAP technology,” ACM SIGMOD Record, pp.65-74, 1997.