Application of Genetic Algorithm for Feature Selection in Optimisation of SVMR Model for Prediction of Yarn Tenacity

A proposed hybrid genetic algorithm (GA) approach for feature selection combined with support vector machines for regression (SVMR) was applied in this paper to optimise a data set of fibre properties and predict the yarn tenacity property. This hybrid approach was compared with a noisy model of SVMR that used all the data set of fibre properties as input in the prediction. The GA for feature selection was used as the preprocessing stage that aimed to find and select the best attributes or variables that most effect or are related to the prediction of yarn tenacity. The hybrid approach showed better predictive performance than the noisy model. However, the results indicated the suitability of GA for feature selection in the choice of the best fibre property attributes that give the preferred performance and high accuracy in the prediction of yarn tenacity.