JOURNAL

Layanan journal yang disediakan oleh Perpustakaan Universitas Gunadarma

Working with Support Vector Regression: Prices Prediction

Judul Artikel:Working with Support Vector Regression: Prices Prediction
Judul Terbitan:Optima Jurnal Keilmuan Aplikasi Teknik dan Manajemen industri
ISSN:0216-0048
Bahasa:IND
Tempat Terbit:Surabaya
Tahun:0000
Volume:Vol. 2 Issue 3 0000
Penerbit:Institut Teknologi Sepuluh Nopember Jurusan Teknik Industri- Fakultas Teknologi Industri
Frekuensi Penerbitan:4 Bulan sekali
Penulis:Budi Santosa
Abstraksi:In this paper, a novel prediction approach, Support Vector Regression (SVR), is utilized to estimate prices. The results are compared to two other methods: Bayesian Neural Networks (BNN), and linear regression (LR). The methods are applied on Boston Housing Price and 'Albuquerque House Price which are nonlinear. The performance of each method is evaluated using mean squared error (MSE) and coefficient of determination R2. Data pre-processing is also investigated to see the impact to the performance error. The results show that SVR performs better than the two other methods. The results confirm that in the case of nonlinear prediction, SVR outperforms traditional regression
Kata Kunci:Jaringan syaraf tiruan; deteksi; generalisasi kesalahan; jaringan saraf feedforward; fungsi kernel; MATLAB; dukungan vektor mesin; setidaknya persegi dukungan vektor regresi; analisis kinerja; statistik; pelatihan sampel
Lokasi:P125
Terakreditasi:belum