Forecasting SET50 index with multiple regression based on principal component analysis

In this paper, we forecast SET50 Index (The stock prices of the top 50 listed companies on SET (Stock Exchange of Thailand) by using multiple regression. At the same time, we examine/ consider/describe? the existence of a high correlation (the multicolinearity problem) between the explanatory variables. One of the approaches to avoid this problem is the use of principal component analysis (PCA). In this study, we employ principal component scores (PC) in a multiple regression analysis. As can be seen, 99.4% of variation in SET50 can be explained by all PCAs. Accordingly, we forecast SET50 Index closed price for the period 1/03/2011 through 31/03/2011 by using three models. We compare loss function, the model forecast explained by all PCs have a minimum of all loss function. -- forecasting ; SET50 index ; multiple regression analysis ; principal component analysis