Implementation of the K-Nearest Neighbor (kNN) Method to Determine Outstanding Student Classes

Abstract

Education being one factor supporting students / I to be able to
increase their knowledge. Each student has their own potential that they have
obtained in the world of education. Therefore, every school has created an
education program that functions to increase the potential of high achieving
students. The program is a flagship class program. What is meant by a
superior class program is a process of selecting and classifying students to be
placed in the classroom superior (grade student / I achievement). Therefore,
this study aims to implement classification on student data using the KNearest Neighbor (kNN) algorithm. K-Nearest Neighbor (kNN) is a method
used to classify data based on training data (data set). The data that the writer
will use is student data of 60 student data. In this classification using the kNN
method aims to classify data on students who are eligible to enter the superior
class (class of outstanding students). The first step is the process of
determining data requirements. Then cleaning or pre-processing and the next
is to design a widget model of the kNN method on the orange application to
carry out the data classification process. The test results using 60 student data
using the KNN method and using the Confusion Matrix obtained an
Accuracy value of 91.6%, then a Precision value of 89.2% and a Recall value
of 92.5%. The conclusion is that this study succeeded in obtaining a method
that the best and also get the best results for Classification of superior student
classes.

GS Cited Analysis