Natural Gas Price Forecasting Research Based On Market Economics

With the crude oil resource becoming less and less, experts foretell that 21 century is the century of Natural Gas. China have entered WTO for two years and have to open some energy districts. Natural gas which is one of the last kind of natural monopoly resources that are priced by government will be gradually changed to be priced by market. Let the price of gas go freely based on western developed countries' experience on controlling natural gas. The world attaches more and more importance to forecast of natural gas price. The research in this thesis is based on the situation in China, which is a forward-looking research.Firstly, we have summarized the research on international and domestic gas price, and at the same time we summed up the methods and models of this forecast. And the several facts which influent the price of gas was carefully analyzed, and abstracted some important factors and then brought forward the model this thesis refers to. Secondly, we discussed theory and arithmetic of partial least squares regression and theory and arithmetic of forecast aspect of BP artificial neural network, thirdly we synthesized the merits of partial least squares regression and theory of non-linear neural network and formed a partial least squares regression and neural network model, using data in natural gas price system in USA market to fit that model, and using acquired data that we have already known to make emulate-forecast and get the error to check over the model's accuracy, we use the model to analyze and emulate present natural gas price in China and make a conclusion that Chinese government have made too low gas price. Finally, we draw a conclusion and acquire some advice.This thesis has useful research and probe on gas price and methods of forecasting it. it has made research achievements as follows:a. It brings forward and apply PLSNN method to forecast and model for the first time based on theory of natural gas economics and deep discussion and analysis on facts which influent natural gas price.b. It created ameliorate arithmetic and program for PLSNN model and validated its high accuracy.c. It uses American data to fit the model and then uses the model to analyze Chinese natural gas price, and brings forward advice on management of Chinese natural gas price.This thesis starts with a wish that Chinese oil and gas exploitation is more reasonable and more scientific. Price is the vane of policy, so we must have a hold on gas price to avoid the mistakes in making policies. We expect that we can find a effective way through our research to forecast natural gas price more accurately, so decision-makers can make better decisions.