H Company’s Order Prediction Research Based On Dynamic Artificial Neural Network

Order forecast is the blood of the enterprise life, it helps us understand the market demand, timely analysis of market information, on the other hand it helps the enterprises to develop effective marketing strategy to provide support. Neural network has been widely used in various fields of natural science and social science, and has achieved a lot of results. In the e-commerce background, market has put forward higher requirements in demand forecasting and order processing speed and precision.This thesisfocuses on the order forecasting of H foreign trade company. Firstly, this thesis analyzes the neural network and order forecasting research status in domestic and foreign;Secondly,the least squares methodis discussed as the theoretical basis of the proposed dynamic artificial neural network model.Thirdly,a dynamic neural network(DNN) model with the comparable optimization ability is proposed based on artificial neural network.The training process of function optimization nonlinear algorithm process is improved by least squares method. Compared to the BP neural network structure,this model is more simpleand the data utilization is also higher.Finally, we make a study on H company about its organization structure and order forecasts status. Some data are collected and analyzed.The order is predicted usingthe proposed DNN.The application results show that the proposed DNN model is effective, faster, more accurate. We alsofind there are still more room for H company’s growth in innovative products and services, management level, order forecasting and management. Therefore some management implications are put forward which including focusing on market research and market study, maintain focusing product, and learning global culture.This research provides a valid orders prediction method for foreign trade companies.