Application of support vector machine and genetic algorithm optimization for quality prediction within complex industrial process

Yarn quality prediction plays an important role in modern textile production management. Due to the nonlinearity and non-stationarity of yarn quality indicator series, the accuracy of the commonly used conventional methods, including regression analyses and artificial neural networks (ANN), has been limited. A prediction model based on support vector regression (SVR) is proposed in this paper to solve the yarn quality prediction problem. Model selection which amounts to search in hyper-parameter space is performed for study of suitable parameters, C and σ, with real code Genetic Algorithms (RGA). The predictive powers of the RGA-SVM models are estimated by comparison with ANN models. The experimental results indicate that in the small data sets and real-life production, the RGA-SVM models have the stability of predictive accuracy, and more suitable for noisy and dynamic spinning process.