TY - JOUR
TI - Bidirectional analysis model of green investment and carbon emission based on LSTM neural network
AU - Hu Yiguo
JN - Thermal Science
PY - 2023
VL - 27
IS - 2
SP - 1405
EP - 1415
PT - Article
AB - Clarifying how the green investment alleviates carbon emissions paves the way for achieving carbon peak and carbon neutralization at a faster pace. In order to propose an effective evaluation model and analyze the interaction between green investment and total carbon emissions, we first and foremost collected data from 30 provinces in China from 2007 to 2019. Secondly, we introduced long short-term memory (LSTM) neural network model, with the amount of government investment in pollution control and environmental infrastructure construction as the model input variables. We also select the total amount of carbon emissions as the model output variables to obtain a neural network model with multiple inputs and a sin-gle output, which can effectively analyze the potential relationship between green investment data and the total amount of carbon emissions data. Then, the OLS model is introduced to test the relationship obtained by LSTM neural network model and analyze its robustness. As a result, the experiment indicates that the LSTM network conceived by us has reliable robustness and fitting performance, with green investment positively affecting total carbon emissions. Meanwhile, we give corresponding policy recommendations according to the model results.