CD Tesis
Analisis Regresi Logistik Ordinal Dan Pendeteksian Outlier Untuk Identifikasi Bakat Olahraga
The purpose of this study is to explain the techniques in detecting the outlier of the ordinal data by using the regression coefficients. Identification data derives from the sport tests of Junior High Schools Students (SMP) in Pekanbaru. The ordinal logistic regression is combined with the proportional odds model. As the attraction of the modulus ordinal regressions interest, the jackknife technique is applied. The satisfactory results show through the graphically outlier identification. The average observation of the regression coefficients appears the same impact that against the regularly best application deviance observation. Eventually, the absolute average of the regression coefficients can be used as an alternative technique in identifying the outlier existence and the residual deviances.
Keywords: outlier, ordinal, proportional odds model, regression coefficients, goodness of fit test model, jackknife
Tidak tersedia versi lain