218
reads

Qiuyan Xu

2009-03-11
10:20:00 - 12:00:00

Linear Regression Analysis Extension and Application in Financial Market

405 , Mathematics Research Center Building (ori. New Math. Bldg.)

In this talk, we will briefly discuss the common problems in classic linear regression models when the data are from financial and economic markets. Heteroskedasticity and serial correlations of the regression residuals call for the extension of the linear regression model. We will introduce the Generalized Least Square method in parameter estimation. We will also discuss a real-life example in detail. Daily return data of 10 companies in the computer sector from Jan.3rd, 2001 to Dec. 30th, 2005 will be used. An illustration on R will be presented, including summary statistics and plot, regression model fit and diagnostic.