失业风险预警系统研究

A Study on the Pre-warning System for Unemployment Crisis

基于我国29年的历史数据,采用BP神经网络模型对我国的失业风险预警问题进行研究。结果表明,神经网络方法在失业风险系统中具有优良的预警效果,其对失业风险综合警情值的预测误差小于3%。相对于景气分析预测法、时间序列分析、灰色预测模型以及回归预测模型等技术,神经网络方法不仅具有良好的预测精度,同时还具备较强的容错能力和泛化能力。因此,在构建我国的失业风险预警系统中,神经网络模型应该是一种被优先考虑的方法。

Based on the economic data of 29 years the research focuses on China＇s unemployment crisis pre-warning issue using the BP neural network model. The research finding indicates that the neural network method has high precision in the unemployment crisis pre-warning with the mean error rate is less than 3 percents. Compared to the Business Forecast Model, Time Series Analysis, Gray Forecast method and Regression Analysis method, the neural network have good fault tolerant and generalization ability. So, we hold that the neural network method is claims precedence over all others method in making Unemployment Crisis Fore-alarming System.