Double Hurdle Model for Censor Data

Abstract

Double hurdle model is used to describe the relationship between the response variables and predictor
variables. Here, the response variable in censored data and predictor variables in continuous, discrete, or a mixture of
both. Censored data have zero value for partly observations, and the rest have varying value. Another its characteristic
is partly values of specific range are transformed as a single value. The classical linear regression analysis can not be
used to see the relationship of variables that are censored. Meanwhile, the logistic regression analysis can not describe
the observation values vary. Therefore, the double hurdle approach can be used to resolve this issue. Double Hurdle
Model is a combination of logit and truncated regression.
Keywords: Censored Data, Double Hurdle