The design of early warning software systems for financial crises in high-tech businesses using fusion models in the context of sustainable economic growth.

Abstract

Enterprises are urged to continue implementing the sustainable development strategy in their business operations as "carbon neutrality" and "carbon peak" gradually become the current stage's worldwide targets. High-tech businesses (HTE) need to be better equipped to manage financial risks and avoid financial crises in the face of severe market competition. The most popular machine learning models-logistic regression, XGBoost, and BP neural networks-are chosen as the base models in this study. The three models are combined using the stacking method to train and forecast the fusion models while offering other researchers some basic model research ideas. The financial crisis early warning (FCEW) of HTE is built concurrently by contrasting the fusion of various quantitative basis models and the fusion procedures of voting and averaging. The outcomes demonstrate that the fusion model outperforms the single model in terms of performance, and the stacked fusion model has the best early warning impact. By comparing and comparing the effect of three fusion models on financial crisis warnings of high-tech enterprises, it makes up for the defect of low accuracy of traditional forecasting methods. It improves the sustainable development path of enterprises.