Improving for the Model of Demand Selfadaption Optimization

The paper forecasts natural gas demand by applying linear regression forecast model, artificial neural network forecast model, grey system forecast model. To integrated utilize information offered by forecast models, avoid dropping useful information by a model, decrease randomness and increase accuracy, which the paper applies optimization combination forecast model and self-adapt optimization combination forecast model to forecast natural gas demand and obtains better results. But the paper finds weight coefficients of self-adapt optimization combination forecast model which is presumed a constant of moved sample data. The method does not embody forth time effect of data. Based far-smallness near-bigness theory of informa-tionism, the paper firstly import time effect function to improve model for natural gas demand selfadaption optimization combination forecast. As a result, it is suggested that the model is better than others.