Risma Moulidya Syafei, Devi Ajeng Efrilianda

Machine Learning Model Using Extreme Gradient Boosting (XGBoost) Feature Importance and Light Gradient Boosting Machine (LightGBM) to Improve Accurate Prediction of Bankruptcy

Introduction

Machine learning model using extreme gradient boosting (xgboost) feature importance and light gradient boosting machine (lightgbm) to improve accurate prediction of bankruptcy. Improve bankruptcy prediction accuracy using Machine Learning. This study implements LightGBM optimized with XGBoost Feature Importance on Taiwanese bankruptcy data, achieving 99.23% accuracy.

Abstract

Abstract. Humans have limitations in processing and analyzing large amounts of data in a short time, including in terms of analyzing bankruptcy data. Bankruptcy data is one of the data that has complex information, so it requires technology that can assist in the process of analyzing and processing data more quickly and efficiently. Data science technology enables data processing and analysis on a large scale, using parallel processing techniques. Parallel processing can be implemented in machine learning models. Purpose: Using parallel processing techniques, data science technologies enable data processing and analysis at scale. Parallel processing can be implemented in machine learning models. Therefore, this study aims to implement a machine learning model using the Light Gradient Boosting Machine (LightGBM) classification algorithm which is optimized using Extreme Gradient Boosting (XGBoost) Feature Importance to increase the accuracy of bankruptcy prediction. Methods/Study design/approach: Bankruptcy prediction is carried out by applying LightGBM as a classification model and optimized using the XGBoost algorithm as a Feature Importance technique to improve model accuracy. the dataset used is the Taiwanese Bankruptcy dataset collected from the Taiwan Economic Journal for 1999 to 2009 and has 6,819 data. Taiwanese Bankruptcy is unbalanced data, so this study applies random oversampling. Result/Findings: The results obtained after going through the model testing process using the confusion matrix obtained an accuracy of the performance of LightGBM+XGBoost Feature Importance of 99.227%. Novelty/Originality/Value: So it can be concluded that the implementation of XGBoost Feature Importance can be used to improve LightGBM's performance in bankruptcy prediction.

Review

The paper presents a timely and relevant study addressing the critical need for efficient and accurate bankruptcy prediction, a task often challenging due to the large volume and complexity of financial data. The authors propose an interesting machine learning approach that leverages the strengths of two prominent gradient boosting algorithms: Light Gradient Boosting Machine (LightGBM) for classification, optimized by the feature importance capabilities of Extreme Gradient Boosting (XGBoost). This novel combination aims to overcome human limitations in data analysis, promising enhanced prediction accuracy, which is highly valuable for financial risk management and decision-making. The methodology outlines a clear strategy, employing LightGBM as the primary classifier while utilizing XGBoost's feature importance to presumably refine the input feature set for LightGBM, thereby improving its performance. The study uses the Taiwanese Bankruptcy dataset, spanning from 1999 to 2009 and comprising 6,819 instances, acknowledging its imbalanced nature and addressing it through random oversampling. While the reported accuracy of 99.227% is remarkably high, especially for a complex task like bankruptcy prediction, a more comprehensive evaluation using metrics beyond just overall accuracy, such as precision, recall, F1-score, or AUC-ROC, particularly for the minority class (bankrupt companies), would provide a more robust assessment of the model's true predictive power and generalizability, given the potential for overfitting with high accuracy on oversampled imbalanced data. The core conclusion that XGBoost Feature Importance significantly improves LightGBM's performance in bankruptcy prediction is a compelling finding, highlighting a potentially effective synergy between these algorithms. The claimed novelty lies in this specific optimization approach, which could offer practical benefits in real-world financial applications. To further strengthen the contribution, future work could involve comparing this integrated model against other state-of-the-art bankruptcy prediction methods, conducting rigorous cross-validation, or providing insights into which specific features were deemed most important by XGBoost, thereby offering interpretable insights into the drivers of bankruptcy. Overall, the paper introduces a promising technique that merits further exploration and validation in the domain of financial distress prediction.

Full Text

Comments

You need to be logged in to post a comment.