Generation of cotton fiber length probability density function with length measures

• Published in: Journal of the Textile Institute Vol. 103; no. 2; pp. 225 - 230
• Main Authors: Lin, Qian, Xing, Mingjie, Oxenham, William, Yu, Chongwen
• Format: Journal Article
• Language: English
• Published: Manchester Routledge 01.02.2012; Taylor & Francis Ltd
• Subjects: Cotton; Economic models; fiber length measures; finite mixture model; length statistics; Probability; probability density function; stepwise regression; Studies
• ISSN: 0040-5000; 1754-2340
• DOI: 10.1080/00405000.2011.566005

This report describes an investigation into the generation of fiber length distributional function using a limited number of length measurements. Based on the experimental data, the correlation between fiber length statistics (the calculated mean value and variance) and length measures (tested short fiber content, span lengths, etc.) was identified by employing the stepwise regression analysis, and this was necessary to describe the location and dispersion of the fiber length distribution. Then, a finite mixture model was adopted to generate the probability density function of fiber length. This paper concludes with a presentation of the approach to determine the probability density function of cotton fiber length. It is shown that the calculated short fiber contents and upper-quartile length from the probability density function fit the experimental values well.