A Probabilistic Prediction Method of Carbon Financial Default Risk Based on Python and Machine Learning

The increasing prominence of environmental problem has aroused the attention of scholars around the world, in order to deal with global warming change, countries have accelerated the construction of carbon market, 2020, our country put forward the "double carbon" goal strategy, under the guidance of the "double carbon" goal, carbon financial trading is in full swing, however, there are problems in the predicting process of carbon financial risks: The first is the diversity of data sources, the integration of enterprise basic information, energy, environment, enterprise operation data and other multi-source data. The second is because China's carbon finance market is in the early stage and the standard of carbon finance has just been released. At present, relatively few enterprises participate in carbon finance and not many enterprises have credit default, making it difficult to obtain the overall sample data. In view of the above problems, based on the existing risk prediction methods, this paper sorts out two kinds of calculation indicators suitable for the risk prediction of a single enterprise, and explains the verification method of the accuracy of the forecast results. On this basis, the predictive process automation is realized by using Python and machine learning.