Study on Improved PCA-SVM Model for Water Demand Prediction

Abstract

construct an improved water demand prediction model for support vector machine (SVM) on the basis of principle components analysis (PCA) in order to improve the accuracy of water demand prediction and prediction efficiency. Analyze the principal components of all the index factors which affect water demand; eliminate redundant information between the indices, thus to reduce SVM input dimensions; besides, it also introduces genetic algorithm, solved the problem that the traditional SUV parameters cannot optimized dynamically. A simulated experiment proves that the predication accuracy of this model is higher than SVM, BP neural network; this model has higher generalization ability and is an effective model for predicting water demand.