Investigation on Stiffness of Finished Stretch Plain Knitted Fabrics Using Fuzzy Decision Trees and Artificial Neural Networks

Stiffness is one of the most important utility properties of textile materials and plays a significant role in well-being due to its influence on physiological comfort [1]. On that point are a great deal of structural properties of textile materials also operating parameters (knitting+finishing) inf...

Full description

Saved in:
Bibliographic Details
Published in:Fibers and polymers 2021, 22(2), , pp.550-558
Main Authors: Baghdadi, Rania, Alibi, Hamza, Fayala, Faten, Zeng, Xianyi
Format: Article
Language:English
Subjects:
Artificial neural networks
Chemistry
Chemistry and Materials Science
Computer architecture
Customer satisfaction
Decision trees
Engineering Sciences
Fabric analysis
Fabrics
Knitting
Neural networks
Parameters
Polymer Sciences
Prediction models
Stiffness
섬유공학
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Stiffness is one of the most important utility properties of textile materials and plays a significant role in well-being due to its influence on physiological comfort [1]. On that point are a great deal of structural properties of textile materials also operating parameters (knitting+finishing) influencing stiffness and there are also statistically significant interactions between the principal factors determining the stiffness of textile materials. As part of our research, we proposed to facilitate the industry adjust the most relevant operating parameters before actual manufacturing to reach the desired stiffness and satisfy consumers. It warrants the application of artificial neural nets (ANNs) to predict the stiffness of finished knitted fabrics and the utilization of the Fuzzy Decision Tree in the selection procedure, to puzzle out the problem of insufficient data and boil down the complexity of predictive models. Moreover, a virtual leave one out approach dealing with overfitting phenomenon and allowing the selection of the optimal neural network architecture was applied.
ISSN:1229-9197
1875-0052
DOI:10.1007/s12221-021-9314-8