PENGKAJIAN PEMBENTUKAN MODEL KLASIFIKASI DALAM PENGELOMPOKKAN JURUSAN SISWA DI SMA (Studi Kasus: Siswa SMA Negeri Siau Timur Kabupaten Siau Tagulandang Biaro Propinsi Sulawesi Utara )

Nelda Ponto (SMA Siau Timur Kabupaten Siau Tagulandang)

MATEMATIKA DAN PEMBELAJARAN
Vol 4, No 1 (2016): MATEMATIKA DAN PEMBELAJARAN

Modeling that involve categorical response variables give important role in the classification problem. Statistical analysis is applied to solve this problem are discriminant analysis and multinomial logistic regression. Implementation of both methods against student of senior high school of East Siau data produce multinomial logistic regression as best method for classify the students into Scicence Program, Social Program, and Language Program. Classification accuracy of model from resampling is 88.1% and of model validation from Tagulandang Senior High School is 70.6%. The variables give significantly effect in classification students to Science Program or Language Program are Mathematics, English, Chemistry, and German, whereas, classification students into Social Program or Language Program are Economy, English, German, and History. Key words: Discriminant analysis, classification, multinomial logistic regression