Hourly Water Demand Forecast Model Based on ν-Support Vector Machine

As BP neural network suffers drawbacks like the choice of the topology structure and overfiting,an hourly water demand forecast model based on ν-support vector machine(ν-SVM) was developed on the basis of the correlation analysis of the hourly water demand series.In ν-SVM algorithm,a parameter ν was introduced to replace the insensitive parameter e of traditional SVM algorithm to effectively control the number of support vector.Case study shows that the modeling speed of ν-SVM-based hourly water demand prediction model is faster and the forecast precision is higher than those of BP neural network-based model and SVM-based model.