Two-part model with measurement error. Modello a due parti con errore di misura

In many applications, there are positive-valued phenomena which show a very high frequency at zero. One major difculty with this type of data is that the existence of a point-mass at zero makes common distributions unsuited for modeling the data. To cope with these difculties, some models have been developed. A popular example is the two-part model in which two stochastic models are assumed: the rst governs whether the response variable is zero or positive an the second, conditional on its being positive, models the level. We extend the two-part model to cope with measurement error on the dependent variables of both stochastic parts. This situation is common in many applied works.