Research on Credit Score Model of Suppliers of Power Grid Enterprises

Aiming at the business needs of power grid companies to score their suppliers' credit, this paper uses logistic regression algorithm as a technical means to build a credit scoring model for power grid companies' suppliers.In this paper, data cleaning and feature extraction are first performed on the training data set of power grid enterprise suppliers. Afterwards, it is trained by the LogisticRegression class in the machine learning library sklearn. The test results show that the detection accuracy of this test is 0.78. The construction of the credit scoring model is suitable for the industry characteristics of power grid companies. However, there is a problem that the credit score is inaccurate due to the lack of authority and uniformity of the evaluation criteria in the training data.