Ethos
Vol 7 No.2 (Juni, 2019) Ethos: Jurnal Penelitian dan Pengabdian kepada Masyarakat (Sains & Teknologi

Penerapan Linier Discriminant Analysis pada Klasifiksi Distress Binsin Perbankan

Asep Nana Rukmana (Universitas Islam Bandung, Bandung, Jawa Bandung, Indonesia)
Bambang Siswoyo (MA’SOEM UNIVERSITY, Jatinangor, Jawa Barat, Indonesia)



Article Info

Publish Date
28 Jun 2019

Abstract

Abstract. Linear discrimant machine learning analysis is part of artificial intelligence that can learn from past data, recognize patterns to get optimal solutions. The prediction of the bankruptcy of a Sharia public bank company in Indonesia is very important. Modeling Machine Learning with five input-output models can be implemented between financial ratio variables against bankruptcy. Overall, a linear discrimant analysis algorithm is able to train data to build patterns of input-output relationships and modeling behavior well. Every company certainly wants an appropriate and efficient decision making. Linear discrimant analysis builds predictive models using financial ratio variables as predictors. In this study the model can recognize well the pattern of financial ratios with the results of model validation in the form of means square error 8% and coefficient terminated 98%.Abstrak. Linier discrimant analisys machine learning merupakan bagian dari kecerdasan buatan yang dapat belajar dari data masa lalu, mengenali pola untuk mendapatkan solusi yang optimal. Prediksi kebangkrutan suatu perusahaan bank umum Syariah di Indonesia sangat penting. Modeling Machine Learning dengan lima model input-output dapat diimplementasikan antara variabel rasio keuangan terhadap kebangkrutan. Secara keseluruhan, algoritma linier discrimant analysis mampu melatih data untuk membangun pengenalan pola hubungan input-output dan perilaku pemodelan dengan baik.  Setiap perusahaan tentu menginginkan sebuah pengambilan keputusan yang tepat dan efisien. Linier discrimant analisis membangun model prediksi menggunakan variabel rasio keuangan sebagai prediktor. Pada penelitian ini model dapat mengenal dengan baik pengenalan pola rasio keuangan dengan hasil validasi model berupa means square error 8% dan koefesien diterminasi 98%.

Copyrights © 2019






Journal Info

Abbrev

ethos

Publisher

Subject

Education

Description

Academic journals that adhere to the highest standards of peer review. The journal is centered in the research and dedication in the science and technology field. The character of the journal are open and contribute to the results of research and dedication and represent of guiding belief in ...