基于LightGBM的信息技术上市公司财务危机预测研究

Research on Financial Crisis Prediction of Information Technology Listed Companies Based on LightGBM

摘要: 信息技术行业技术更新快，竞争非常激烈信息技术行业特有的风险因素，如创新要求高、融资渠道狭窄、偿债能力较差、现金流量不稳定。为了监督信息技术类上市公司运营状况，警惕退市风险，避免财务危机的发生，本文在国泰安数据库收集了深交所与上交所在A股上市的472家信息技术类公司的公开信息，基于LightGBM集成算法构建了信息技术类上市公司的财务危机预测模型，并对其特征重要性进行排序，研究发现，研究收入占营业收入的比例对信息技术上市公司是否会面临退市风险最重要，除此之外，本文还对模型进行了网络参数调优，调优后的模型AUC值0.923，接近于1，模型预测效果好。

Abstract: The information technology industry features a rapid pace of technological updates and extremely fierce competition. There are also risk factors unique to this industry, such as high requirements for innovation, narrow financing channels, relatively poor debt repayment abilities, and unstable cash flows. In order to monitor the operating conditions of listed information technology companies, stay vigilant against delisting risks, and avoid the occurrence of financial crises, this paper collected the public information of 472 information technology companies listed on the A-share market of the Shenzhen Stock Exchange and the Shanghai Stock Exchange from the GTA (Guotai’an) Database. Based on the LightGBM integrated algorithm, a financial crisis prediction model for listed information technology companies was constructed, and the importance of its features was ranked. The research findings show that the proportion of research income to operating income is the most crucial factor in determining whether information technology listed companies will face delisting risks. In addition, this paper also carried out network parameter optimization for the model. After optimization, the AUC (Area Under the Curve) value of the model is 0.923, which is close to 1, indicating that the model has a good prediction effect.