Left truncated regression model is a regression model which the parameters prevailed for the dependent variable values y > a, with a is a left truncation point that is choosen according to the experiment goals, and a is also the observation object itself at once. Thus, truncated regression model is more correct if it is used for the experiment that oriented on a specific characteristic of the observation object, that is dependent variable. The distribution that is used for this regression model is normal left truncated distribution. Estimation of regression parameters used Maksimum Likelihood method with Newton’s iteratif method.
Copyrights © 2006