Fecha de publicación: 29/09/2023 Fuente: Wipo "BigData" The invention relates to the technical field of credit risk assessment, in particular to a method for carrying out credit evaluation on a market subject based on government affair big data, comprising the following steps: carrying out government affair data fusion government; establishing an index system; extracting features; extracting labels; constructing a credit evaluation model; the beneficial effects are that the method for performing credit evaluation on the market subject based on government affair big data is based on national and provincial enterprise credit information sharing and collection, performs fusion treatment on data of each department, applies a semantic recognition technology, performs structuralization on government affair text information, deeply mines related texts, extracts risk control features and labels, and performs credit evaluation on the market subject based on the government affair big data. The problem that the number of features and labels is insufficient is solved, the basis of intelligent algorithm application is tamped, the enterprise credit evaluation model is constructed by applying logistic + ensemble learning, the credit score and the enterprise default probability are linked, and the application performance of the enterprise credit score in the bank credit risk control decision process is improved.