Early Detection of Financial Distress at PT. Garudafood Putra Putri Jaya Tbk Using Machine Learning Algorithms

Abstract

This study aims to develop an early warning system to detect potential financial distress at PT Garudafood Putra Putri Jaya Tbk using Machine Learning (ML) algorithms. A quantitative case study approach was applied using the company’s publicly available quarterly financial reports from 2019 to 2024. Five ML algorithms were tested: Logistic Regression, Random Forest, Naive Bayes, K-Nearest Neighbors (KNN), and Support Vector Machine (SVM). The Random Forest algorithm achieved the highest accuracy of 85.71%, with a precision of 100%, recall of 66.67%, and an F1-score of 80%. Operating Profit Margin and Debt to Asset Ratio were identified as the most significant indicators in detecting financial distress. The proposed model can serve as a decision-support tool for management, investors, and regulators to proactively anticipate financial problems and reduce bankruptcy risk through early intervention. Keywords: Abstract, aim, scope, result