POSSIBILITY LINEAR REGRESSION ANALYSIS WITH A NEW METRIC

In this paper, we propose a new method for computation of fuzzy regression the method is constructed on the metric that minimizing the total distance between the observed prediced values. This method is powerful to estimate the parameters of fuzzy numbers are influenced on the estimations. Also we use possibility function in constrains. To compare the performance of the proposed approach with the previos method, two examples are presented.

Keywords:

P-ideal , implicative ideal , positive implicative ideal