Modeling Research On Demand Forecasting Of Excavator

With the implementation of Chinese large-scale infrastructure construction as well as the "The Belt and Road" strategy,China construction machinery industry has made considerable progress.However,behind the rapid development of China construction machinery industry,We are still facing the problems of oversized rate and amplitude of market demand change,overcapacity.The fundamental reason is that there is no scientific and accurate market demand forecast in the industry,resulting in the contradiction between production and demand.Taking excavator products as an example,a variety of mathematical models have been established in this paper after the analysis of the influencing factors.And good prediction results have been achieved in practical applications.This paper first summarizes the research status of demand forecasting methods at home and abroad.Combined with the qualitative analysis of the excavator market,five factors that affect the market demand are summarized,such as investment of capital construction,national macro-control,needs of depreciation and replacement of equipment,profitability of customers and seasonal factors.For monthly demand,the Winters addition/multiplication model and SARIMA model are established in this paper.By comparing the results,the actual prediction results of the model SARIMA(0,1,1)(2,1,0)is better.Through the market research,a more objective and accurate calculation method of the excavatorownership is obtained,and the feasibility of the grey model and multiple linear regression model in the prediction of excavator ownership is verified.This paper also provides a new method to predict the annual demand of the excavator.By analyzing the factors affecting customers' purchase of excavators,the model of the customers' profitability structure is established.The proportion of customers'profitability structure is taken as a time series,and the prediction performance of the Elman neural network is verified.Finally,based on the proportion of the customers' profitability structure,the BP neural network model is established.The prediction results show that the prediction accuracy is obviously better than the multiple linear regression model.