Research on AI-Based Financial Intelligent Decision-Making System

This paper presents the design and implementation of an enterprise financial intelligent decision making system based on Python programming language and SpringBoot framework. The system uses the XGBoost algorithm for credit scoring and adopts a two-layer LSTM network to construct a risk warning model. The system integrates real-time data acquisition, multi-dimensional analysis and intelligent alerting capabilities. A historical dataset of financial records from 3,000 companies was used to train and test the system. The findings indicate an F1 score of 0,856 for risk alerts and an AUC value of 0,912 for credit scoring. The system is deployed on a distributed architecture and has an average response time of 185 ms with 200 simultaneous users. In practice, the system has significantly improved the efficiency of the company's financial decision making by reducing analysis time to 3 minutes, increasing alert lead time to 12 days and increasing annualised ROI by 2.8 percentage points.