Research of Tax Revenue Intelligent Forecast System

The work of tax revenue forecast is very important. Current revenue forecast system in our country has just begun, so compared with foreign country, there is a larger gap. Theutility of tax forecast system is still relatively low-level for decision-making reference. In this paper, we propose an intelligent forecast system. By means of advanced machin elearning methods, via establishing of more scientific predictive analysis model on non-linear data, the system can process noisedata, and realize real-time forecast and model learning on line, as well as establish a prediction, analysis model with learning function. So the model can be adjusted timely according to the data in history or just recently emerged, to improve prediction accuracy, and to predict the data unable to be collected for directing our work. Via the factors in great events the system learns to predict the impact analysis.