World Carbon Emissions Prediction Model And Applications Based On Hierarchical Cluster And BP Neural Network

Global warming is a man- made phenomenon, not only harm the natural balance of the ecosystem, but also a threat to human survival and development. Intergovernmental Panel on C limate C hange(IPCC) Fourth Assessment report pointed out that global warming is mainly carbon dioxide generated by the use of fossil fuels caused. Forecast of world carbon emissions led to concerns about global warming has become a hot issue of academic research. World Carbon Emissions Forecasting Model is an important content, generally use single prediction model, integrated model predictions. Two prediction models have advantages and disadvantages, on the basis of previous studies, integrated forecasting model for the non-depth study, made world emissions of carbon prediction model fusion system clustering and BP Neural Network.Firstly, the two different world's carbon emissions forecast models to summarize and compare the pros and cons of various models. The results showed that a single prediction model is simple to operate, objective and credible results, but many assumptions and forecast accuracy is not high; fusion prediction model is relatively complex, but the method is flexible, widely used, information utilization, and high prediction accuracy is relatively high.Secondly, the complexity and diversity of the world's carbon emissions forecast indicators, forecasting model to build the world's carbon emissions system clustering and BP neural network, and describes the applicability of the model in world carbon emissions forecast.Finally, from the economic factors, social factors, environmental factors and other four point of view, select 10 indicators of per capita GDP, the total population of the forest area, fossil fuel energy consumption, the use of cluster analysis method in SPSS to extract the full impact on the world carbon emissions projections indicators in order to build a different world carbon emissions predictor layer, so that the world's carbon emissions predicted value close to the actual value of the world's carbon emissions considerably. The design good indicator layer input MATLAB software for BP neural network comp uting, the corresponding predictions obtained by using the prediction model compared with each other to determine the optimal index layer, and in 2015, in 2016 the world's carbon emissions were 400.01( PPM), 402.98(PPM).In this study, experimental verification of the world's carbon emissions forecast model system clustering and BP neural network based on accurate higher than the traditional single BP neural network prediction model has better applicability.