REGRESI LINEAR TERSENSOR STUDENT-T; STUDENT-T CENSORED LINEAR REGRESSION

ERISKI ISNANDA, Zulaela

2015 | Skripsi | FMIPA

Tobit regression is a regression analysis that address the issue of data censored cases where the error is assumed to be normally distributed. But for the case of data with an error that is not normal, this tobit regression produce biased estimates. So use Censored Linear Regression Student-T where this regression is error assumed to be Student-T distributed. Censored Linear Regression Student-T, because the degree of distribution parameters Student-T provides good dimensional to achieve robust statistical inference. This illustrates censored regression models using a different approach than the Tobit models, are quite susceptible to the presence of outliers.

Kata Kunci : Tobit Regression; Linear Regression Censored Student-T, Degrees of Freedom parameter