Ridge Regression for Estimation of Transition Probabilities from Aggregate Data

Inderdeep Kaur, M. B. Rajarshi · Communications in Statistics - Simulation and Computation · 2011

When data from several independent Markov chains are aggregated over each time point, least square estimation of transition probabilities faces the problem of multi-collinearity. We propose here an estimation procedure which involves use of ridge regression for the ordinary least square estimators. Performance of this estimator is then compared with that of the ordinary least squares.

Read the paper · More papers on PaperTik