Analysis of Financial Risk Early Warning on the Chinese Internet Companies via Machine Learning

Han Yating

Abstract

In recent years, the rapid development of internet and artificial intelligence technologies has prompted a growing number of traditional enterprises to pursue transformation and upgrade. Despite the numerous benefits that internet-based enterprises offer, including low cost, high efficiency, and low risk, the Internet economy also introduces new risks. Chinese internet-listed companies, in particular, have been subject to increasingly stringent restrictions imposed by the country's financial market, rendering their market environment more complex while increasing their operational risks. To address these challenges, we employed machine learning-based risk warning models to predict potential financial crises and to help Chinese internet-listed companies prevent such risks more effectively. Specifically, we trained and tested a series of models using the financial data of 387 Chinese internet-listed companies from 2018 and 2019, with the aim of predicting their financial conditions in 2020. The resulting predictions were then compared to gauge the efficacy of the models. Our findings indicate that the overall prediction accuracy of the models exceeded 90%, far surpassing that of traditional single-variable warning models. This comparative analysis provides guidance and suggestions for companies seeking to select appropriate financial risk warning models, thereby affording our study valuable reference value.