Generating cotton fiber length probability density function from fiber length parameters by finite mixture model

Abstract

This paper generated the fiber length probability density function of different cotton samples on the basis of Finite Mixture Model. With the mixture of two Weibull distributions, the commonly used and discrete fiber length parameters were employed to obtain the function. Curve fitting of fiber length frequency histograms was carried out and the differences between the model proposed by this paper and previous model were compared. The results show that when fiber length measures were modified by the related linear relational expressions, this model, compared with other models, could fit the test data very well with the errors all less than 0.0045.