Estimating Regression Parameters with Imprecise Input Data in an Appraisal Context

Traditional valuation models, such as capitalization or comparative, share the common principle of using exact information in their calculations. However, in the professional field, it is difficult to find situations characterized by the wealth and precision of the information available. Within the professional context, it is usually necessary to carry out estimates on some of the explanatory variables of pricing, even though consciously introducing an important degree of subjectivity. The model proposed in this paper, which uses the goal programming optimization technique, is capable of dealing with imprecise information, since it allows to consider intervals to define the values of the explanatory variables. Additionally, this paper presents and demonstrates several propositions that increase the informative capacity of the model presented.