Prediksi Financial Distress Pada Perusahaan Sektor Energi Di BEI Menggunakan Random Forest Dan Support Vector Machine

Qolbi, Nur Latifatul (2026) Prediksi Financial Distress Pada Perusahaan Sektor Energi Di BEI Menggunakan Random Forest Dan Support Vector Machine. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221102-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Sektor energi Indonesia menghadapi tekanan signifikan akibat fluktuasi harga komoditas global, terutama setelah krisis energi dan konflik geopolitik pada 2022, yang berdampak pada menurunnya kinerja keuangan sejumlah perusahaan dan meningkatkan risiko financial distress. Penelitian ini bertujuan memprediksi financial distress pada perusahaan sektor energi periode 2022–2024 serta mengidentifikasi rasio keuangan yang paling berpengaruh. Pendekatan pelabelan untuk menghasilkan label distress yang digunakan adalah K-Means clustering yang berbasis pola data intrinsik, Hasil pelabelan itu kemudian dijadikan target pada pemodelan machine learning menggunakan algoritma Random Forest dan Support Vector Machine (SVM). Untuk mengatasi ketidakseimbangan data, digunakan metode Synthetic Minority Over-sampling Technique (SMOTE) pada data training. Penelitian ini menunjukkan bahwa mayoritas perusahaan tergolong non-distress, di mana metode K-Means mengidentifikasi 15,3% (33 perusahaan) sebagai distress dengan tren menurun. Kemudian hasil dari perbandingan metode Random Forest dan Support Vector Machine berdasarkan kinerja 5 metrik evaluasi, didapatkan bahwa model Random Forest menghasilkan kinerja paling baik dan stabil dalam memprediksi financial distress perusahaan sektor energi di BEI dengan akurasi 95,45%, presisi, sensitivitas, dan F1-score 85,71%, serta AUC 96,91%. Variabel paling berpengaruh dalam memprediksi financial distress berdasarkan model terbaik yaitu rasio profitabilitas, khususnya Return on Equity (ROE), Return on Assets (ROA) dan Net Profit Margin (NPM).
=================================================================================================================================
The Indonesian energy sector faces significant pressure due to global commodity price fluctuations, particularly following the energy crisis and geopolitical conflicts in 2022, which have impacted the financial performance of several companies and increased the risk of financial distress. This study aims to predict financial distress in the energy sector during 2022–2024 and identify the most influential financial ratios. Labeling approaches used to generate distress labels include K-Means clustering based on intrinsic data patterns and the Ohlson model as a theory-based approach. These labels were then used as targets in machine learning modeling using Random Forest and Support Vector Machine (SVM) algorithms. To address data imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied to the training data. The results indicate that most companies are classified as non-distress, with K-Means identifying 15.3% (33 companies) as distress, showing a decreasing trend over time. A comparison of Random Forest and SVM based on five evaluation metrics reveals that Random Forest performs best and most stably in predicting financial distress in BEI-listed energy companies, achieving an accuracy of 95.45%, precision, sensitivity, and F1-score of 85.71%, and an AUC of 96.91%. The most influential predictors in the best-performing model are profitability ratios, specifically Return on Equity (ROE), Return on Assets (ROA), and Net Profit Margin (NPM).

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Financial Distress, K-Means, Random Forest, Sektor Energi, Support Vector Machine, Financial Distress, Energy Sector, K-Means, Random Forest, Support Vector Machine
Subjects:	H Social Sciences > HG Finance > HG4028.V3 Valuation. Economic value H Social Sciences > HG Finance > HG4529 Investment analysis Q Science > QA Mathematics > QA353.K47 Kernel functions (analysis) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) Q Science > QA Mathematics > QA9.58 Algorithms Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Nur Latifatul Qolbi
Date Deposited:	02 Feb 2026 02:42
Last Modified:	02 Feb 2026 02:42
URI:	http://repository.its.ac.id/id/eprint/131384

Actions (login required)

View Item