PEMILIHAN MODEL REGRESI LINEAR TERBAIK BERDASARKAN MODIFIKASISTATISTIK Cp MALLOWS (Studi Kasus : Faktor-Faktor yang Mempengaruhi Indeks Prestasi Mahasiswa D3 MI F MIPA UNS)

YUNIATI, TINA (2010) PEMILIHAN MODEL REGRESI LINEAR TERBAIK BERDASARKAN MODIFIKASISTATISTIK Cp MALLOWS (Studi Kasus : Faktor-Faktor yang Mempengaruhi Indeks Prestasi Mahasiswa D3 MI F MIPA UNS). Other thesis, Universitas Sebelas Maret.

[img] PDF - Published Version
Download (320Kb)

    Abstract

    . There are several methods on determining of best linear regression model, i.e the forward selection, the backward elimination procedure, stepwise regression procedure, and all possible regression. All possible regression method is frequently used. This has several criteria, adjusted R 2 , S 2 and Mallows Cp. Mallows Cphas connection with residual sum of square (SSE), and residual mean square (MSE). MSE is frequently used to estimate 2 s . Since the distribution of Cpwhich is founded from the complete model gives the value of Cp ¹ p, so it is used modified Cp (Cp*) . That has E(Cp * ) = p . The purpose of this research is to find the best regression model of the factors that affect the IP values of D3 student MIF MIPA UNS ( UAN value, the value of subjects and support facility ) based on modification statistics of Cp Mallows. The result of discussion indicate that influence IP value of D3 students MI FMIPA UNS is value of subjects, i.e. Basic of Management, Introduction of Computer Science and Basic of Mathematics, with linear regression model is Y= 1.19 + 0.210 X3 + 0.192 X7+ 0.220 X9 Key word : best linear regression model, Mallows Cp

    Item Type: Thesis (Other)
    Subjects: Q Science > Q Science (General)
    Q Science > QA Mathematics
    Divisions: Fakultas Matematika dan Ilmu Pengetahuan Alam
    Fakultas Matematika dan Ilmu Pengetahuan Alam > Matematika
    Depositing User: Andika Setiawan
    Date Deposited: 18 Jul 2013 11:22
    Last Modified: 18 Jul 2013 11:22
    URI: http://eprints.uns.ac.id/id/eprint/6700

    Actions (login required)

    View Item