Supervised Learning Approach for Predicting the Quality of Cotton Using WEKA

Cotton is the world's most important natural fibre used in Textile manufacturing. Cotton fiber is processed into yarn and fabric. Yarn strength depends extremely on the quality of cotton. The physical characteristics such as fiber length, length distribution, trash value, color grade, strength, shape, tenacity, density, moisture absorption, dimensional stability, resistance, thermal reaction, count, etc., contributes to the quality of cotton. Hence determining the quality of cotton accurately is an essential task to make better raw material choices in textile industry which in turn will support better buying and selling decisions. In this work, cotton quality prediction is modeled as classification task and implemented using supervised learning algorithms namely Multilayer Perceptron, Naive Bayes, J48 Decision tree, k-nearest neighbor in WEKA environment on the cotton quality assessment dataset. The classification models have been trained using the data collected from a spinning mill. The prediction accuracy of the classifiers is evaluated using 10-fold cross validation and the results are compared. It is observed that the model based on decision tree classifier produces high predictive accuracy compared to other models.