improving time series forecastingaccuracy is an important yet often difficult task.both theoretical and empirical findings haveindicated that integration of several models is an effectiveway to improve predictive performance, especiallywhen the models in combination are quite different. in this paper,a model of the hybrid artificial neural networks andfuzzy model is proposed for time series for...

Support vector regression (SVR) is well known as a or prediction tool under the Machine Learning (ML) which preserves all key features through training data. Different from general prediction, here, we proposed SVR to predict new approximate solutions after generated some iterates using an iterative method called Lanczos algorithm, one class of Krylov solvers. As know that solvers, including me...