This paper first introduces the basic concepts related to interval data,and then summarizes the classical methods of dealing with interval data by predecessors.The author developed a new method of processing interval data with polynomials.The process is to use the polynomial regression models to deal with the interval median and interval radius of dependent variables.Also the classical algorithm of matrix de-composition in numerical algebra is introduced to simplify the calculation,because this process involves matrix inversion.In view of the limitation of positive interval radius,the author adjusts the results of regression analysis with mathematical algorithm,and finally returns the prediction of dependent variables.Through Monte Carlo simulation analysis and empirical research on a set of bi-ological data,the author verifies the feasibility of the new method,and compares it with traditional methods.It is found that the error of data prediction is smaller and the method is effective. |