Prediction of financial failure in informatics sector with machine learning methods

Information technology companies have great importance in the rapidly evolving technology sector today. Operating in the fields of computers, software, telecommunications, the internet, and other digital technological services, IT companies play a pioneering role in innovation, data processing, communication, and digital transformation. These companies have a significant impact on the business world, creating value-added products and contributing to the country's economy. However, companies can also experience periods of failure. When a company's financial performance is not at the expected or desired level, it can be considered as a case of failure. This situation is assumed to have a negative impact on the company's revenues, profitability, cash flow, debt situation, and other financial indicators. Financial failure can adversely affect a company's sustainability, growth potential, and competitiveness. The financial failure of IT companies is related to factors such as competition in the industry, technological changes, customer demands, costs, management errors, and other factors. Therefore, predicting and preventing financial failure in advance is important for ensuring the sustainability of companies, managing financial risks, and developing appropriate strategies. In this regard, machine learning methods provide a powerful tool for predicting financial failure by using large datasets. These methods help predict the likelihood of financial failure for companies by analyzing financial data, identifying patterns, and determining risk factors. These predictions can assist companies in taking early action, developing appropriate strategies, and reducing the risk of financial failure. This study utilized Decision Trees, Logistic Regression, Artificial Neural Networks, Support Vector Machine, Random Forest, and Naive Bayes methods to predict financial failure based on the financial statements of 29 IT companies listed on BİST (Borsa Istanbul). The results showed that machine learning methods are effective tools for predicting financial failure. The Random Forest method stood out with the highest accuracy value (97.52%), while the Support Vector Machine method also achieved successful results (94.30%). The accuracy values of the other methods were as follows: Decision Tree (89.48%), Artificial Neural Networks (90.79%), Logistic Regression (82.75%), and Naive Bayes (66.81%). These findings demonstrate that it is possible to predict the financial failure of IT companies in advance and that machine learning methods are valuable tools in this regard. This study encourages further research to support the strategic decisions of IT companies, reduce financial risks, and ensure their sustainability.