E-commerce Turnover Prediction Based On Genetic Algorithm Optimized Grey Neural Network

As a main part of the new economy development,E-commerce changes rapidly in recent years.The development goal of China's E-commerce is to fully integrate E-commerce into all areas of the national economy,and accumulate E-commerce transactions to 40 trillion RMB in 2020.E-commerce transaction volume is an important manifestation of E-commerce's development level,and it has become an inevitable requirement to provide a basis and reference for government and enterprises' decisions by using scientific and reasonable ways in forecasting E-commerce transaction volume.In the background of E-commerce's rapid development,scholars begin to analyze and forecast the development of E-commerce since 2003.In the aspect of data selection,since lack of historical data,the content of the research is mainly focused on the local quantization,questionnaire design and so on.In terms of the calculation method,the weighted sum method,the analytic hierarchy process,and fuzzy evaluation method are very prevalent.Starting from introducing the background and current situation of the electronic commerce,this paper constructs a forecasting index system of E-commerce transactions based on domestic and foreign electronic commerce development research results,and then uses various intelligent algorithms on forecasting China's E-commerce transactions for empirical research purpose.This research mainly includes the following three parts:The first part is about the forecasting index construction for E-commerce turnover.Through analyzing a large number of literatures and collecting related factors of E-commerce transactions,this paper divides various factors into three dimensions based on qualitative analysis,and then by using gray correlation analysis method and quantitative analysis,this paper integrates key factors within three different dimensions and builds a forecasting index for E-commerce transaction volume.The second part is about the establishment of the forecasting model for E-commerce transactions.With introducing the basic theory of grey system,artificial neural network algorithm,genetic algorithm,this paper successively establishes BP neural network model.Through analysis,the shortcomings of these three models are found out.The improved grey neural network model after genetic optimization is used to predict the amount of E-commerce transactions.The third part is the empirical study on forecasting E-commerce transactions.By selecting the data of E-commerce transactions in China from 2005-2014,this paper uses the above model to forecast the amount of E-commerce transactions in China,and compared the stability and accuracy of each model.The results show that using genetic algorithm optimized grey neural network for E-commerce turnover prediction is better than other models.