Prediction And Analysis Of China-ASEAN Foreign Direct Investment Based On Multi-model Fusion

With the development of the world regional economy,foreign direct investment has become a dominant force that cannot be ignored.Studying the trend of foreign direct investment in China and ASEAN countries not only enriches the theoretical research on foreign direct investment,but also meets the needs of economic analysis of China-ASEAN Ocean Data Platform,and provides reference for the governments of China and ASEAN countries in formulating policies and strategies on foreign direct investment.Until now,most of the research on foreign direct investment is based on traditional methods.With the maturity of machine learning and artificial intelligence technology,good research results have been achieved in many fields.Therefore,this paper combines machine learning methods to explore and study the trend of foreign direct investment has certain theoretical significance and practical value.In this paper,two foreign direct investment prediction models are proposed based on whether the foreign direct investment feature data of China and ASEAN countries in the China-ASEAN Ocean Big Data Platform Investment Database is complete.Firstly,in view of the completeness of foreign direct investment feature data,this paper proposes a multi-model fusion foreign direct investment forecasting model based on feature analysis,called MP-RBP-FDI prediction model.This prediction model uses the MP algorithm to deal with the feature of foreign direct investment,and combines the advantages of random forest algorithm and BP neural network.By comparing with the single MLR,RF and BP neural network prediction models,the MP-RBP-FDI prediction model has improved in both prediction accuracy and generalization ability.Secondly,in view of the lack of foreign direct investment feature data,this paper proposes a multi-model fusion foreign direct investment prediction model based on time series analysis,called ALS-FDI prediction model.The model is based on the ARIMA model and the LSTM network.Comparing with the single GM(1,1),ARIMA and LSTM network prediction models,the proposed ALS-FDI prediction model effectively solves the problem of both linear and nonlinear components in the foreign direct investment data series,and has higher accuracy than other prediction models.Finally,the two prediction models proposed in this paper are applied to the China-ASEAN Ocean Big Data Platform.Users can select sub-module based on feature data or historical data to check the predicted value according to whether the feature data of foreign direct investment is complete or not,which fully reflects the practical value of the research results in this paper.