Prediksi Financial Distress Perusahaan Sektor Industri di Indonesia dengan Metode Klasifikasi dan Melibatkan Synthetic Features Generation

Muda, Muhammad Adlansyah (2021) Prediksi Financial Distress Perusahaan Sektor Industri di Indonesia dengan Metode Klasifikasi dan Melibatkan Synthetic Features Generation. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
06211740000074-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only

Download (1MB) | Request a copy

Abstract

Masalah kondisi financial distress dapat berakhir dengan kebangkrutan apabila tidak segera ditanggulangi. Untuk mengantisipasi dan meminimalisir dampak dari bangkrutnya suatu perusahaan terutama pada sektor industri, maka dilakukan prediksi financial distress untuk menilai kondisi keuangan perusahaan dan perspektif masa depannya. Pada penelitian ini prediksi financial distress dilakukan dengan metode klasifikasi seperti Generalized Extreme Value Regression, Logistic Regression, Support Vector Machine, dan Extreme Gradient Boosting dengan melibatkan synthetic features generation secara serentak dan seleksi variabel. Berdasarkan nilai accuracy, AUC, dan F1-score dari hasil evaluasi model menggunakan data testing didapatkan bahwa metode synthetic features generation tidak selalu memberikan performansi klasifikasi terbaik pada tiap size. Pada size 0 dan size 1 disimpulkan bahwa model Extreme Gradient Boosting dengan melibatkan synthetic features generation dan seleksi variabel merupakan model dengan performansi klasifikasi terbaik, sedangkan pada size 2 dan size 3 didapatkan bahwa model Extreme Gradient Boosting tanpa melibatkan synthetic features generation dengan seleksi variabel merupakan model dengan performansi klasifikasi terbaik dalam memprediksi kondisi keuangan perusahaan sektor industri di Indonesia. ==================================================================================================================== The problem of financial distress can lead to bankruptcy if it is not addressed immediately. To anticipate and minimize the impact of a company bankruptcy, especially in the industrial sector, financial distress predictions are made to assess the company’s financial condition and its future perspective. In this study, prediction of financial distress is carried out using classification methods such as Generalized Extreme Value Regression, Logistic Regression, Support Vector Machine, and Extreme Gradient Boosting by involving synthetic features generation simultaneously and variable selection. Based on the accuracy, AUC, and F1-score from the results of model evaluation using data testing, it is found that the synthetic features generation method does not always provide the best classification performance for each size. At size 0 and size 1, it can be concluded that the Extreme Gradient Boosting model involving synthetic features generation with variable selection is the model with the best classification performance, whereas at size 2 and size 3, it is found that the Extreme Gradient Boosting model without involving synthetic features generation with variabel selection is the model with the best classification performance in predicting the financial condition of industrial sector companies in Indonesia.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Drift, Financial Distress, Klasifikasi, Sektor Industri, Synthetic Features Generation, Classification, Industrial Sector
Subjects: H Social Sciences > HD Industries. Land use. Labor > HD108 Classification (Theory. Method. Relation to other subjects )
Q Science > Q Science (General) > Q325.5 Machine learning.
Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression
T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User: Muhammad Adlansyah Muda
Date Deposited: 05 Mar 2021 00:23
Last Modified: 05 Mar 2021 00:23
URI: https://repository.its.ac.id/id/eprint/83474

Actions (login required)

View Item View Item