利用獨立成份分析的功能，解決求解投組分配的困難，再用LAVE,GARCH,跟RiskMetrics 三種不同的變異數方法去配適獨立成份的動態過程，並利用台股指數進行一天的風險值預期，共一千天，最後用回顧測試檢定模型的優劣

The Value at Risk (VaR) measures the potential loss in value of risky asset or portfolio over a defined period for a given confidence interval. The traditional way needs to estimate corresponding distribution and process of portfolio, which is very difficult. Independent component analysis (ICA) is designed for detection of blind folded signals and retrieves out of a high-dimensional time series stochastically independent source components. We can use the property of independence to estimate distribution of portfolio easily. This paper uses three different volatility estimate methods in conjunction with independent component process to calculate value at risk.

[1] S.D. Campbell, A review of backtesting and backtesting procedures, Journal of Risk 9, pp.1–17, 2006
[2] J. Cardoso ,High-order Contrasts for Independent Component Analysis, Neural Computation11(1), 157-192.1999
[3] J.F Cardoso, Dependence, correlation and Gaussianity in independent component analysis, The Journal of Machine Learning Research, v.4 n.7-8, p.1177-1203, October 1 - November 15, 2004
[4]R.J Carroll, and D. Ruppert, Transformation and Weighting in Regression, Chapman and Hall, New York, 1988
[5] S. M. Cha and Laiwan Chan, Applying Independent Component Analysis to Factor Model in Finance, Intelligent Data Engineering and Automated Learning - IDEAL 2000, Data Mining, Financial Engineering and Intelligent Agents, ed. K.S. Leung, L.W. Chan and H. Meng, Springer, pp 538-544, 2000
[6] Y. CHEN, W. Härdle, S.O. Jeong, Nonparametric risk management with generalized hyperbolic distributions. SFB 649, discussion paper 2005-001.
[7] Y. CHEN, W. HÄRDLE, V. SPOKOINY, Portfolio value at risk based on independent components analysis, J. Comput. Appl. Math, 205 pp. 594—607,2007
[8] P. Comon, Independent Component Analysis: a new concept?, Signal Processing, Elsevier, 36(3):287--314 (The original paper describing the concept of ICA),1994
[9] Duffee and Pan, An Overview of Value at Risk, Journal of Derivatives, Spring 1997
[10] A. Hyvärinen, NewApproximations of Differential Entropy for Independent ComponentAnalysis and Projection Pursuit, MIT Press, Cambridge,MA, pp. 273–279.1998
[11] A. Hyvärinen, Independent Component Analysis: ATutorial CIS, Helsinki University of Technology, Finland,April, 1999.
[12] A. Hyvärinen, E. Oja, Independent component analysis: algorithms and applications, Neural Networks 13 411–430.1999
[13] A. Hyvärinen, J. Karhunen, E. Oja, Independent Component Analysis, Wiley, NewYork, 2001.
[14] O. V. LEPSKI, A problem of adaptive estimation in Gaussian white noise. Theory Probab. Appl. 35 454-466,1990
[15] O. V. LEPSKI, E. MAMMEN, and V. POKOINY, Optimal spatial adaptation to inhomogeneous smoothness: An approach based on kernel estimates with variable bandwidth selectors. Ann. Statist. 25 929-947, 1997
[16] L. D. Lathauwer, An introduction to independent component analysis, 14:123–149, 2000
[17] D. Mercurio, V. Spokoiny, Statistical inference for time-inhomogeneous volatility models, Ann. Statist., 577–602, 2004
[18] V. SPOKOINY, Estimation of a function with discontinuities via local polynomial fit with an adaptive window choice. Ann. Statist. 26 1356-1378, 1998