Composite Prediction Method Based On Quasi-linear Regression And Its Application

The prediction work, to discuss the future development of things, in recent decadeshas caused more and more attention of the people. All of the management decisions shouldbe based on the forecast for the future. If it is able to make an accurate enough predict onthe future development of the things, there is no doubt that it can provide a basis forpeople to make reasonable decision. Thereby we can reduce the occurrence of errors andobtain a better effect.Regression analysis, as an important branch of statistics, is an effective tool forscientific prediction. Regression model is a very classic single equation econometricmodel. People have been constantly perfected the existing models and discover newprediction methods, such as the combination forecast method, etc. But the methodimprovement when the data size is too large needs to be improved. This paper mainly usedthe theoretical basis of the regression model. On the basis of it this paper applied thequasi-linear regression model, and put forward the composite prediction method.Genetic algorithm, one of evolutionary algorithms, is an optimization searchalgorithm in computational mathematics. In this paper, a new regression model namedquasi-linear regression model was given some detail introduction. And we used geneticalgorithm for its solution. Composite prediction method puts forward the idea of piecewiserandom sampling based on the existing prediction model. It can process the data whosesize is very large, and then choose the right prediction model. Composite predictionmethod in a certain extent solved the burden of prediction model due to the large data sizeand gave a method for determining the accuracy of the results. And that has been verifiedwith an example.