With the growth of financial data scale, the traditional financial anomaly identification method is difficult to meet the demand for efficient monitoring. This study constructs an intelligent anomaly identification model for financial data based on decision tree algorithm, which improves the accuracy and efficiency of financial anomaly detection. The study utilizes variance reduction and information gain as decision criteria, adopts post pruning strategy and advancement technique to optimize the model performance, and combines a variety of data preprocessing methods to process financial data containing 4616 listed companies in 26 industries. The experimental results show that the constructed model achieves an accuracy rate of 0.869, a precision rate of 0.830, a recall rate of 0.791, and an F1 value of 0.810 in financial data anomaly recognition, which is better than the comparative algorithms such as CNN, LSTM, and BP. In the financial fraud identification task, when both corporate governance indicators and financial control indicators are used, this model achieves an accuracy rate of 0.884 and an AUC value of 0.835, which is an improvement of 7.3% to 22% over the comparison models. The case study verifies the effectiveness of the model in identifying financial anomalies, and achieves accurate identification of anomalous signals such as abnormal net profit per capita and high equity pledge. The study shows that the decision tree-based intelligent anomaly identification model for financial data can accurately extract key features in financial time series data, improve the accuracy of anomaly identification, and provide support for the early warning of corporate financial risks.
