Predicting thermal resistance of cotton fabrics by artificial neural network model

•Thermal resistance of handloom cotton fabrics has been predicted using artificial neural network models.•Four basic fabric constructional parameters namely EPI, PPI, warp count and weft count were taken as the inputs.•Eight network structures having 5–12 nodes in the hidden layer were tried for predicting thermal resistance.•Best prediction results were obtained for ANN model with 7 nodes in the hidden layer.•Ranking of input parameters was determined by the input saliency test of the ANN model.

This paper presents the prediction of thermal resistance of handloom cotton fabrics by artificial neural network models using four primary fabric construction parameters, i.e. ends per inch (EPI), picks per inch (PPI), warp count and weft count as the inputs. ANN model with seven nodes in the single hidden layer exhibited the overall best performance with coefficient of determination of 0.90 and 0.86 and mean absolute error of only 5.13% and 4.23% during training and testing respectively. The importance of fabric construction parameters on the thermal resistance of fabrics was also analyzed by the developed ANN model. Weft count, EPI and warp count were found to be the first three most important fabric constructional parameters in descending order of importance in predicting thermal resistance of plain woven cotton fabrics.