Continuity of response variable is an important assumption in probabilistic linear regression models. So, these models cannot be used for cases in which response variable is not continuous. One common method in such circumstances is implementation of of an important In i) exact observed data, ii) exact relationship between variables. But, in practice, observations of some variables and/or relationships among the variables might be vague. Such situations may be occurred as a result of inexact report of response variable and/or expert discretion because of his/her uncertainty about the existence of a characteristic. In such cases, fuzzy models can be used for modelling and predicting.
