Using Random Forest Algorithm in financial Failure Prediction- An Empirical Study

Abstract

Predicting financial failure is a modern topic in financial management. It serves as an early warning to detect future failures, and thus avoid them by taking the necessary actions by decision-makers.

The current research aims to highlight the most important modern techniques that banks can use to detect the likelihood of their financial failure in the future. To achieve this goal, the Random Forest algorithm were used as one of the most important modern techniques based on artificial intelligence, along with a number of traditional models. The Sherrord (1987), Kida (1981), and Zmijewsk (1984) models were used, and then a comparison was made between the traditional and modern methods to diagnose which of the two methods provides more accurate information in judging the continuity of banks, as well as studying the actual reality of the last two years of the banks under investigation.

The research community represents all banks operating in the Iraqi banking sector. The research sample included two private Iraqi banks that were selected based on their continued operation during the research period, which included ten years, starting from (2013) until (2022). The data were obtained by using the financial reports published by the banks in the research sample, in addition to the Iraq Stock Exchange website.The research employed a deductive approach for the theoretical framework and an applied approach for the practical application. After data collection, the researcher processed and prepared the data for use in traditional prediction models. These models were then used as inputs for the Random Forest and Nearest Neighbor algorithms. Several computer programs were utilized, including Microsoft Excel for data entry and Python for writing the algorithms and extracting their results.

The research reached a number of conclusions, the most important of which is that the use of AI-based prediction techniques, represented by the random forest algorithm, provides more accurate and reliable predictions that help in judging the ability of banks to continue operating compared to other traditional techniques.