Eearly-warning of China's cotton production is very useful to cotton's supply demand balance

cotton price's stable and cotton policy's consistence. It should be progressed in early-warning's indicatorspicking and forecasting precision. This study uses systematic scienti?c approach

considering ?ve aspects of cotton supply

demand

economic environment

price and production

select 15 indicators and construct the early warning index system of cotton production early warning system in China

and a grey correlation analysis was used to evaluate the relevance of the index system. Based on ARMA model and AMRA-BP network

an early warning model of cotton production in China is established. The result shows that the model has a greater improvement than the existing cotton production early warning research in terms of index selection and early warning accuracy

and both models get good prediction results

but the accuracy of the combined model is much higher than the single model

it is more suitable for such forecasting issue.