Financial Risk Prediction of Manufacturing Enterprises Based on SMOTE-Ensemble Learning

Abstract

In the era of big data and artificial intelligence, the traditional financial risk prediction methods can not meet the actual needs of enterprises. “Intelligence + Finance” is the main force of financial development in the new era. Based on this, this paper selects the manufacturing industry which has an important supporting position in the national development as the research object, and uses smote algorithm to over sample the data to solve the problem of data imbalance. Then, by comparing and analyzing the prediction results of different machine learning models, XGBoost, Bagging, KNN and Random Forest are selected as the base models to construct the ensemble learning model. The research indicates that the prediction accuracy of the ensemble learning model is as high as 98.08%, and is better than that of the single model. This finding can provide a new financial risk prediction path for manufacturing enterprises, help enterprises to predict financial risk more efficiently, and promote the sustainable and stable development of enterprises.