Analysis and Prediction of Air Permeability of 100% Cotton Single Jersey Fabric Using Machine Learning Approaches

Authors

  • Sharmin Akter Daffodil International University
  • Sohel Ahammed
  • Md. Mashiur Rahman Khan

Abstract

The air permeability property has great significance on the fabric comfort related property. In this study Artificial Neural Network (ANN), Random Forest and Additive Regression Classification Models have been applied for the prediction of the air permeability of single jersey knitted fabrics made of 100% cotton fiber. For this aim, 100 different single jersey knitted fabrics were used and there basic properties such as yarn linear density, tightness factor, fabric loop length, fabric thickness, stitch density, fabric unit weight, and air permeability properties were evaluated. ANN, Random Forest, and Additive Regression Classification Models were developed to predict air permeability properties of single jersey fabrics. It was found that all the models give outputs closer to the experimental results. However, ANN estimation success was found higher than other models.

Downloads

Published

2020-06-12

Issue

Vol. 11 No. 3 (2020)

Section

Scholarly Articles

License

By submitting a manuscript, the authors agree that the copyright of all materials published in Journal of Textile and Apparel Technology and Management (JTATM) is transferred to the publisher if and when their manuscript is accepted for publication. Please note the following details as an understanding for submitting a manuscript to JTATM for publication.
Authors grant rights to JTATM to disseminate the paper and its contents as a file on the Internet when the paper is accepted. JTATM owns an exclusive right for the paper until it is either rejected or published. The authors cannot submit the same paper in part or in its entirety to another journal during the review.
Authors retain rights to be identified as the authors when the paper is published, and the patent right or rights relating to products or processes described in the paper. The authors retain rights to the educational and research uses of the paper, such as teaching and exchange of the published or pre-publication version of the paper with colleagues, for non-commercial purposes. The authors may obtain permission from the journal for a non-exclusive use of the contents in the paper, such as tables, figures and texts, with a proper acknowledgement to the journal.
Authors must warrant the originality of the paper in that the contents of the paper in part or in full have not been published or submitted elsewhere while the paper is under review.
Effective date for the above policies is December 1, 2010. The authors of the papers published before this date will be contacted via email for copyright transfer agreements.