Wijaksana, Steven (2023) Analisis Perbandingan Klasifikasi Financial Distress Perusahaan Sektor Transportasi di Bursa Efek Indonesia Menggunakan Artificial Neural Network dan Support Vector Machine. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 06311940000001-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
06311940000001-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (2MB) | Request a copy |
Abstract
Financial Distress merupakan kondisi kesulitan keuangan perusahaan yang berujung pada kebangkrutan. Kondisi ini disebabkan oleh beberapa faktor, salah satunya adalah faktor krisis ekonomi. Krisis ekonomi pernah dirasakan beberapa kali di Indonesia, antara lain pada tahun 1997 (Asian Financial Crisis), 2008 (Global Financial Crisis), dan 2020 (Covid-19 Outbreak). Salah satu sektor yang paling terdampak signifikan pada saat krisis adalah sektor transportasi dan logistik. Oleh karena itu, agar investasi dalam sektor ini memberikan imbal hasil maksimal, maka diperlukan pemilihan perusahaan yang sehat secara keuangan, yakni dengan cara melihat indikator rasio keuangan dari perusahaan-perusahaan tersebut. Penelitian berfokus pada pembentukan model Support Vector Machine (SVM) dan Artificial Neural Network (ANN) terbaik melalui data training, yang digunakan untuk mengklasifikasikan variabel respons 6 perusahaan transportasi di Bursa Efek Indonesia ke dalam kategori Financial Distress (FD) atau Non-Financial Distress (NFD) berdasarkan variabel prediktor. Data yang dipakai dalam penelitian merupakan rasio keuangan perusahaan sektor transportasi di Bursa Efek Indonesia pada periode kuartal III 2015 - kuartal IV 2020 (sebagai data training) dan periode kuartal I 2021 - kuartal IV 2021 (sebagai data testing). Sedangkan variabel prediktor dan respons yang dipakai dalam penelitian berturut-turut merupakan 15 nilai rasio keuangan dan 1 variabel klasifikasi biner: FD=1 atau NFD=0. Model SVM terbaik dicari melalui fitur GridSearchCV, sedangkan model ANN terbaik dicari melalui kriteria nilai Mean Squared Error (MSE) terkecil pada data testing. Selain itu, karena ANN dan SVM termasuk supervised learning, maka ditentukan klasifikasi awal perusahaan ke dalam kedua kategori respons menggunakan Altman Z-Score. Nilai accuracy score kedua model selanjutnya dibandingkan untuk menentukan model klasifikasi terbaik. Hasil penelitian menunjukkan model SVM terbaik merupakan model SVM kernel Radial Basis Function parameter C=1 dan γ =1, dengan accuracy score model sebesar 70,83%, sedangkan model ANN terbaik merupakan model MLP (15-4-1), yaitu model Multilayer Perceptron With Backpropagation dengan 15 input neuron, 4 hidden neuron, dan 1 ouput neuron, dengan accuracy score model sebesar 79,17%. Berdasarkan nilai accuracy score, dapat disimpulkan bahwa model ANN terbaik, yaitu MLP (15-4-1), lebih akurat dalam hal klasifikasi dibandingkan model SVM terbaik, sehingga metode ANN dapat menjadi metode alternatif klasifikasi Financial Distress selain metode Altman Z-Score.
==============================================================================================================================
Financial Distress is a condition of a company's financial difficulties that leads to bankruptcy. This condition is caused by several factors, one of which is the economic crisis factor. The economic crisis has been felt several times in Indonesia, including in 1997 (Asian Financial Crisis), 2008 (Global Financial Crisis), and 2020 (Covid-19 Outbreak). One of the sectors largely damaged from crisis is the transportation and logistics sector. Therefore, to gain maximum returns on investment in this sector, investors need to select a healthy company by looking its financial ratio indicators. This study focuses on establishing the best Support Vector Machine (SVM) and Artificial Neural Network (ANN) models from training data, which are later used to classify the response variables of 6 transportation companies on the Indonesia Stock Exchange into Financial Distress (FD) or Non-Financial Distress (NFD) categories, based on their predictor variables. The training and testing data used in this study are successively the financial ratios of transportation sector companies on the Indonesia Stock Exchange in the third quarter of 2015 - fourth quarter of 2020, and in the first quarter of 2021 - fourth quarter of 2021, while the predictor and response variables used in this study successively represents 15 financial ratios and one binary classification variable: FD=1 or NFD=0. The best SVM model is examined through the GridSearchCV feature, while the best ANN model is examined through the least Mean Squared Error (MSE) in the testing data criteria. Moreover, because both ANN and SVM use supervised learning algorithms, the initial classification of companies into both response categories is determined using the Altman Z-Score. The accuracy scores of the two models are then compared to get the best classification model. The results showed that the best SVM model is the SVM model with the parameter C=1,γ =1, and Radial Basis function kernel, with an accuracy score of 70.83%, while the best ANN model is the MLP (15-4-1) model, or the Multilayer Perceptron With Backpropagation model with 15 input neurons, four hidden neurons, and one output neuron, with an accuracy score of 79.17%. Based on the accuracy score, we can conclude that the best ANN model, or MLP (15-4-1), is more accurate than the best SVM model. Thus, the ANN method could be used as an alternative method besides Altman Z-Score in terms of Financial Distress classification.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Accuracy Score, Altman Z-Score, Artificial Neural Network, Financial Distress, Support Vector Machine |
| Subjects: | H Social Sciences > HJ Public Finance Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis |
| Depositing User: | Steven Wijaksana |
| Date Deposited: | 20 Jan 2023 02:13 |
| Last Modified: | 20 Jan 2023 02:13 |
| URI: | http://repository.its.ac.id/id/eprint/95485 |
Actions (login required)
 |
View Item |