PEMODELAN B-SPLINE DAN MARS PADA NILAI UJIAN MASUK TERHADAP IPK MAHASISWA JURUSAN DISAIN KOMUNIKASI VISUAL UK. PETRA SURABAYA

Authors

  • I Nyoman Budiantara Lecturer at Department of Statistics, ITS, Surabaya
  • Fredi Suryadi Staf. at Petra Christian University, Surabaya
  • Bambang Widjanarko Otok Ph.D Student at Department of Mathematics, UGM; Lecturer at Department of Statistics, ITS, Surabaya
  • Suryo Guritno Lecturer at Department of Mathematics UGM, Yogyakarta

:

https://doi.org/10.9744/jti.8.1.1-13

Keywords:

nonparametric regression, B-Spline, MARS, determinant coefficients.

Abstract

Regression analysis is constructed for capturing the influences of independent variables to dependent ones. It can be done by looking at the relationship between those variables. This task of approximating the mean function can be done essentially in two ways. The quiet often use parametric approach is to assume that the mean curve has some prespecified functional forms. Alternatively, nonparametric approach, .i.e., without reference to a specific form, is used when there is no information of the regression function form (Haerdle, 1990). Therefore nonparametric approach has more flexibilities than the parametric one. The aim of this research is to find the best fit model that captures relationship between admission test score to the GPA. This particular data was taken from the Department of Design Communication and Visual, Petra Christian University, Surabaya for year 1999. Those two approaches were used here. In the parametric approach, we use simple linear, quadric cubic regression, and in the nonparametric ones, we use B-Spline and Multivariate Adaptive Regression Splines (MARS). Overall, the best model was chosen based on the maximum determinant coefficient. However, for MARS, the best model was chosen based on the GCV, minimum MSE, maximum determinant coefficient. Abstract in Bahasa Indonesia : Analisa regresi digunakan untuk melihat pengaruh variabel independen terhadap variabel dependent dengan terlebih dulu melihat pola hubungan variabel tersebut. Hal ini dapat dilakukan dengan melalui dua pendekatan. Pendekatan yang paling umum dan seringkali digunakan adalah pendekatan parametrik. Pendekatan parametrik mengasumsikan bentuk model sudah ditentukan. Apabila tidak ada informasi apapun tentang bentuk dari fungsi regresi, maka pendekatan yang digunakan adalah pendekatan nonparametrik. (Haerdle, 1990). Karena pendekatan tidak tergantung pada asumsi bentuk kurva tertentu, sehingga memberikan fleksibelitas yang lebih besar. Tujuan penelitian ini adalah mendapatkan model terbaik mengenai nilai ujian masuk terhadap nilai IPK (Indek Prestasi Kumulatif) mahasiswa jurusan Disain Komunikasi Visual tahun 1999 di Universitas Kristen Petra Surabaya dengan analisis regresi, baik parametrik maupun nonparametrik. Pendekatan regresi parametrik menggunakan regresi linear sederhana, kuadratik dan kubik, sedangkan regresi nonparametrik digunakan B-Spline dan Multivariate Adaptive Regression Splines (MARS). Secara keseluruhan, model terbaik dipilih berdasarkan koefisien determinasi terbesar. Namun demikian untuk MARS, model terbaik dipilih berdasarkan pada GCV, minimum MSA dan koefisien determinasi terbesar. Kata kunci: regresi nonparametrik, B-Spline, MARS, koefisien determinasi.

Downloads

Published

2006-10-11

How to Cite

[1]
I. N. Budiantara, F. Suryadi, B. W. Otok, and S. Guritno, “PEMODELAN B-SPLINE DAN MARS PADA NILAI UJIAN MASUK TERHADAP IPK MAHASISWA JURUSAN DISAIN KOMUNIKASI VISUAL UK. PETRA SURABAYA”, Jurnal Teknik Industri: Jurnal Keilmuan dan Aplikasi Teknik Industri, vol. 8, no. 1, pp. 1-13, Oct. 2006.

Issue

Section

Articles