Un'applicazione del filtro di Kalman alla stima precoce del numero di laureati

Abstract

The paper deals with the problem of forecasting the performance of a group of students on the basis of the results coming from the first exams they are submitted A technique based on the Kalman filter recursive algorithm or, more specifically, on the Bayesian approach to the forecasting problem developed by Harrison and Stevens (1976) is proposed. An application based on the individual records of students in Economics at the the University of Rome "La Sapienza" is proposed and some main results evaluated.