Prediction Model with Interval Data -Toward Practical Applications-

The regression model is one of typical model for predicting some values by analyzing existing numerical data collected in various ways. If data are not crisp numbers, they are usually transformed into numerical crisp values by means of some methods such as quantification method. In companies’ decision making process, collected and referred data usually have uncertainty which sometimes play an important roles for business performance. One of authors have proposed a model deriving predicting values with uncertainty by handling data with uncertainties. In this paper, we review some method for this purpose, then describe the model by applying it to some test cases.