Enhanced Financial Performance Classification Using an Improved ID3 Algorithm with Decision Coherence

Abstract

The evaluation of enterprise financial performance is of great significance to investors, creditors and managers. Through the performance evaluation, we can judge the comprehensive strength of the enterprise and then take corresponding countermeasures. The study mainly quantifies the three financial performance factors of profitability, operating ability and solvency in segments. At the same time, variables are constructed for the statement subjects other than the performance factors. In addition, an iterative binary tree three-generation algorithm model based on decision coordination degree is introduced to classify and measure the financial performance of enterprises. The results show that when the sample size is 300, the construction time of the improved model is only 120ms, and out of 50 financial statements, only 7 financial statements with "poor" profitability are misclassified as "excellent". The average classification accuracy of the improved model is 87.57% and the average area under the curve is 0.8369. In the solvency test, the average classification accuracy of the improved model in the solvency indicator is 84.21%. In the test of operating capacity indicators, the mean value of classification accuracy of the improved model proposed by the Institute in operating capacity indicators is as high as 85.16%. The results show that the classification model proposed by the Institute has the least construction time and the highest classification accuracy. It provides reliable technical support for the rational allocation of corporate asset structure and the effective decision-making of financial institutions.