Sustainable production using a hybrid IPSO optimized SVM-based technique: Fashion industry

Abstract

The fashion industry is one of the most polluting industries in the world. Accordingly, it is crucial to gain a greater understanding of the products in this industry in order to prevent environmental degradation. Due to the rapid rate of transformation and the enormous volume of data, it is critical to develop an appropriate feature selection technique that improves managers' decision-making to predict the production of products with minimal environmental pollution. In this research, we proposed a new feature selection method based on Support Vector Machine (SVM) technique involving more than one environmental parameter, such as waste, pollution and energy. It served to predict the sustainable productions that achieve the highest number of customers and the lowest possible damage to the environment. Furthermore, we intended to optimize the SVM technique in feature selection by adopting the improved form of this technique through the Improved Particle Swarm Optimization (IPSO) algorithm and ultimately adjust the different kernel parameters involved in the SVM technique learning process. The overall objective of this study was to evaluate the performance of SVM in combination with IPSO algorithm and its impact on predicting high demand products and to enable smart decisions for managers to make sustainable production compared to other existing techniques. The results suggested that the accuracy of the test in the newly proposed hybrid method is higher than other similar methods owing to optimization of the parameters of SVM algorithm by IPSO as well as the adoption of kernels corresponding to the environmental parameter in SVM technique learning process.