This [unedited] guest post is by a student in my PSTAT262MC class (background post). Please praise/critique/comment on its quality and importance to you.
igor [thumbnails].jpgIgor Himelfarb says: Exploring Factors Predicting Academic Achievement in College To predict first year college GPA (FYGPA) in college-bound senior, a dynamic regression model was constructed where dependent series were FYGPA and independent series contained three SAT scores (Critical Reading, Math, and Writing) and high school GPA (HSGPA). Due to the structure of the data, mean value instead of the actual values for each variable was extracted at each time point and used in the analysis. The model presumed not to have linear trend or seasonality. A plot of two series (Yt and Xt) against time are presented in Blog 1. The nature of the data seems to support a simple, straight line regression where θt represents the regression parameter vector. A one-step forecast is presented in the attached file. igor_blog2.jpg

blog comments powered by Disqus


18 March 2010