Improved Nonlinear Regression Models And Applications Based On Interval Data

Interval data is most common and widely used,so how to effectively analyze interval data,understand its data structure and laws and predict,has become an important issue in various fields.This paper first makes a comparative analysis of the Parametrized Model and Monte Carlo Method proposed by scholars,and makes up for the lack of comparison of interval data analysis methods,which shows that the Monte Carlo method works better when the interval data presents a linear structure and the number of samples is large.Secondly,the idea of Monte Carlo method is extended to the data of nonlinear structure,combined with the advantages of Center and Range Additive Model,the Monte Carlo Additive model MCAM is constructed:on the basis of using all the information in the interval,we can depict the nonlinear data structure data.At the same time,we carny on the data simulation and the empirical analysis,synthetically shows the validity and reliability of the proposed method,and finally draw the conclusion that the Monte Carlo Additive model has a good performance effect when the internal information of the interval is not clear.It provides new solutions for the application field of interval data.