Dewi, Komang Amelia Komala (2024) Analisis Prediksi Financial Distress Perusahaan Publik Sektor Industri di Indonesia dengan Generalized Extreme Value (GEV) Regression dan GEV-Stochastic Search Variable Selection (SSVS). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003201157-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003201157-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (5MB) | Request a copy |
Abstract
Indonesia memiliki perkembangan pembangunan yang cukup cepat dan dinamis. Hal ini dapat dilihat melalui beberapa indikator, salah satunya yakni keberhasilan pembangunan ekonomi. Sektor industri adalah salah satu sektor yang memiliki peran dalam percepatan pertumbuhan ekonomi. Prediksi financial distress memiliki peran penting sebagai early warning system dalam mencegah risiko kebangkrutan perusahaan. Analisis dilakukan terhadap 111 perusahaan publik sektor industri di Indonesia tahun 2005-2018 dengan menggunakan metode analisis berbasis extreme value yakni Generalized Extreme Value (GEV) Regression dan GEV- Stochastic Search Variable Selection (SSVS). Analisis dengan GEV-Regression membutuhkan data yang konvergen untuk memastikan algoritma optimasi menemukan solusi. Analisis GEV-Regression menggunakan variabel respon ROA dan ICR menunjukkan bahwa model terbaik dalam setiap size windowing yakni dengan menggunakan seluruh variabel prediktor. Akan tetapi, seringkali ditemukan bahwa model yang kompleks memiliki resiko overfitting lebih tinggi. Penelitian ini melakukan pengecekan dengan membagi data menjadi training dan testing dan didapatkaan model terbaik GEV-Regression dalam setiap skema dan size windowing adalah dengan variabel prediktor signifikan. Dalam semua skema analisis, variabel EBITA, STA, NPM, dan Inflasi konsisten menunjukkan signifikansi dalam model terbaik. Dalam metode GEV-SSVS, pemilihan model terbaik didasarkan pada frekuensi kemunculan 5% dari iterasi dan dipilih model dengan nilai DIC terendah. Estimasi parameter didasarkan pada nilai median karena parameter model terpilih cenderung asimetris. Analisis dengan GEV-SSVS memberikan hasil klasifikasi yang baik untuk data training dan testing. Variabel EBITA cenderung konsisten dalam setiap skema dan size analisis. Penggunaan windowing size 1, 2, dan 3 pada penelitian ini digunakan untuk melakukan prediksi.
========================================================================================================================
Indonesia has a rapid and dynamic development. This can be seen through several indicators, one of which is the success of economic development. The industrial sector is one of the sectors that has a role in accelerating economic growth. Financial distress prediction has an important role as an early warning system in preventing the risk of corporate bankruptcy. The analysis was conducted on 111 public companies in the industrial sector in Indonesia in 2005-2018 using extreme value-based analysis methods, namely Generalized Extreme Value (GEV) Regression and GEV- Stochastic Search Variable Selection (SSVS). Analysis with GEV-Regression requires convergent data to ensure the optimization algorithm finds a solution. GEV-Regression analysis using ROA and ICR response variables shows that the best model in each windowing size is to use all predictor variables. However, it is often found that complex models have a higher risk of overfitting. This study checks by dividing the data into training and testing and finds that the best GEV-Regression model in each scheme and windowing size is with significant predictor variables. In all analysis schemes, EBITA, STA, NPM, and Inflation variables consistently show significance in the best model. In the GEV-SSVS method, the selection of the best model is based on the frequency of occurrence of 5% of the iterations and the model with the lowest DIC value is selected. Parameter estimation is based on the median value because the parameters of the selected model tend to be asymmetric. Analysis with GEV-SSVS provides good classification results for training and testing data. EBITA variables tend to be consistent in each scheme and size of analysis. The use of windowing sizes 1, 2, and 3 in this study are used to make predictions.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Financial distress, GEV Regeression, GEV-SSVS, Industri, Financial distress, GEV Regeression, GEV-SSVS, Industry |
| Subjects: | H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation H Social Sciences > HA Statistics > HA31.7 Estimation H Social Sciences > HC Economic History and Conditions H Social Sciences > HC Economic History and Conditions > HC441 Macroeconomics. H Social Sciences > HC Economic History and Conditions > HC79.E5 Sustainable development. (circular economy) H Social Sciences > HG Finance Q Science > QA Mathematics > QA274.7 Markov processes--Mathematical models. |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Komang Amelia Komala Dewi |
| Date Deposited: | 08 Aug 2024 04:25 |
| Last Modified: | 08 Aug 2024 04:25 |
| URI: | http://repository.its.ac.id/id/eprint/114459 |
Actions (login required)
 |
View Item |