Open AccessPREDIKSI POTENSI SISWA PUTUS SEKOLAH AKIBAT PANDEMI COVID-19 MENGGUNAKAN ALGORITME K-NEAREST NEIGHBOR
Author(s) -
Irma Darmayanti,
Pungkas Subarkah,
Luky Rafi Anunggilarso,
Jali Suhaman
Publication year - 2021
Publication title -
jurnal sains dan teknologi
Language(s) - English
Resource type - Journals
eISSN - 2548-8570
pISSN - 2303-3142
DOI - 10.23887/jst-undiksha.v10i2.39151
Subject(s) - confusion matrix , confusion , k nearest neighbors algorithm , covid-19 , mathematics education , value (mathematics) , recall , computer science , artificial intelligence , psychology , machine learning , medicine , cognitive psychology , disease , pathology , infectious disease (medical specialty) , psychoanalysis
The implementation of the PSBB has an impact on all sectors, one of which is education, namely the threat of children dropping out of school. Dropouts explain that every student or student who leaves school or other educational institutions for any reason before finishing school without moving to another school. Early prediction must be done, to prevent many students dropping out of school. The dataset used in this study was taken from students in Junior High School (SMP) in Banyumas Regency. The method used in this study is the confusion matrix and 10-fold cross validation on the K-Nearest Neighbors (KNN) algorithm. The results obtained on the KNN algorithm in predicting the potential for dropout students are 87.4214%, with a precision value of 88.2%, recall 87.4% and F-Measure 87%. Then the results of the accuracy value on the KNN algorithm are categorized as Good Classification