Estimation of Transition Probability Matrix in Continuous Evaluation Process of School Examination

Abstract

This paper is about estimation of transition probability matrix using proportion data available from continuous evaluation process of student’s performance in school final examination. Since the available data do not provide individual transition information over time, the method of Maximum Likelihood fails. Weighted Least Square method is applied to obtain Estimates of transition probabilities using real life data. The transition probabilities are estimated using Iteratively Re-Weighted Least Square (IRLS) method where errors are correlated.