College Students' Employment Situation Forecast Model Research

University graduate employment is the decisive factors of the human resources market balance, this paper based on statistic regression theory, analyzes the influential factors of university graduates employment, grasp the different factors on the employment of college graduates for establishing accurate degree, the influence of the university graduates employment forecast system provides the basis, To establish forecasting model of artificial neural network, college graduates employment, and the results are error analysis. Finally, based on statistic regression theory and artificial neural network model, a new college graduates employment for the prediction model, provides new ideas, and education of talents structural reforms and to provide reference for graduate selection.The statistics regression theory of all factors, its core is to calculate the employment and the factors affecting the regression coefficients. Using multiple linear regression theory, through the analysis of the influence factors of university graduates employment, make many complicated factors associated with the quantitative comparison of employment, according to the degree of correlation degree of employment, determine the impact factors. According to China's national bureau of statistics yearbook and local university of data, this paper determines the four kinds of factors including the general economic situation, industry economy, talent cultivation plan and subjective efforts as the main factors affecting the prediction model.Article established statistics regression and artificial neural network is two kinds of prediction model for university graduates employment. Statistics regression model, the output value and the actual employment model for the average difference for predicting error covariance 5.9%,0.0013, Artificial neural network model, the output value and the actual employment model for the average difference value for predicting error covariance 2.7%, 0.00074.Two kinds of model for optimal error inspection level. In addition, the numerical experiments, due to the statistics regression prediction model of solving large inverse matrix, the computational complexity is adopted in this paper, the (LMS) through iterative algorithm for spend operation will reduce computational complexity, the practical problems, to save time about half.Because of the artificial neural network model is higher than the average error covariance and statistical theory, the theory of above two kinds of choices, combined with the advantages of the prediction model. Using statistics regression theory, main factors affecting judgment, and using artificial neural network model for prediction model, the prediction problem provides a new research methods and tools.The source of the model parameters, establishing a scientific and reasonable model, model solving process is simple, understandable results. In the region, and statistical data can be successfully used for university graduates, the model for prediction research and employment factors affecting the correlation analysis. This is our human resource market of macro-control and the research provides a powerful support and quantitative analysis tool.