Hybrid Genetic Algorithm and Support Vector Machine Performance in Public Fiscal Revenue Prediction

摘要

Local public fiscal revenue (PFR) is an important indicator to measure the level of local economic development. Accurate prediction of local public fiscal revenue can provide theoretical support for governments and relevant departments to make scientific decisions. Firstly, we use the principal component analysis to reduce the dimensions of variables. Secondly, we predict Harbin public fiscal revenue by using hybrid genetic algorithm (GA) and support vector machine (SVM) model. Then we compare the results with a back propagation neural network model and a gray neural network model optimized with the genetic algorithm. The results show that the hybrid genetic algorithm and support vector machine model is more effective for forecasting Harbin public fiscal revenue.