Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/60202
Title: Prediksi Tingkat Keberhasilan Mahasiswa Tingkat I IPB Dengan Metode k-Nearest Neighbor
Authors: Sitanggang,Imas Sukaesih
Purnama, Endang
Faiza, Ninon Nurul
Keywords: Bogor Agricultural University (IPB)
k-nearest neighbor
classification
Issue Date: 2009
Abstract: Analysis of academic data and personal data of first-year students in IPB is necessary to predict the successful of their study in the end of the first year. One of techniques in classification that can be used for completing that task is k-nearest neighbor that will build a classifier. This research aimed to develop classifier to predict the successful of first year students at IPB. The attributes used in this research are selected based on target class-influenced statistic hipotesis test. Chi-square test is implemented for nominal attributes whereas Spearman Rank Correlation Coeficient test is used for selecting the numerical attribute. The result of this research is a classifier with accuracy 52.97%.
URI: http://repository.ipb.ac.id/handle/123456789/60202
Appears in Collections:UT - Computer Science

Files in This Item:
File Description SizeFormat 
Abstract.pdf
  Restricted Access
Abstract332.89 kBAdobe PDFView/Open
BAB I Pendahuluan.pdf
  Restricted Access
BAB I342.92 kBAdobe PDFView/Open
BAB II Tinjauan Pustaka.pdf
  Restricted Access
BAB II530.33 kBAdobe PDFView/Open
BAB III Metode Penelitian.pdf
  Restricted Access
BAB III361.48 kBAdobe PDFView/Open
BAB IV Hasil dan Pembahasan.pdf
  Restricted Access
BAB IV473.56 kBAdobe PDFView/Open
BAB V Kesimpulan dan Saran.pdf
  Restricted Access
BAB V344.77 kBAdobe PDFView/Open
Cover.pdf
  Restricted Access
Cover342.87 kBAdobe PDFView/Open
Daftar Pustaka.pdf
  Restricted Access
Daftar Pustaka332.96 kBAdobe PDFView/Open
G09nnf.pdf
  Restricted Access
full text929.37 kBAdobe PDFView/Open
Lampiran.pdf
  Restricted Access
Lampiran488.11 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.