Open AccessAkurasi Klasifikasi Pengguna terhadap Hotspot WiFi dengan Menggunakan Metode K-Nearest Neighbour
Author(s) -
Raemon S Saljumairi,
Sarjon Defit,
S Sumijan,
Yusma Elda
Publication year - 2021
Publication title -
jurnal sistim informasi dan teknologi
Language(s) - English
Resource type - Journals
ISSN - 2686-3154
DOI - 10.37034/jsisfotek.v3i3.55
Subject(s) - rss , euclidean distance , computer science , k nearest neighbors algorithm , signal strength , wireless , pattern recognition (psychology) , artificial intelligence , telecommunications , operating system
The Current wireless technology is used to find out where the user is in the room. Utilization of WiFi strength signal from the Access Point (AP) can provide information on the user position in a room. Alternative determination of the user's position in the room using WiFi Receive Signal Strength (RSS). This research was conducted by comparing the distance between users to 2 or more APs using the euclidean distance technique. The Euclidean distance technique is used as a distance calculator where there are two points in a 3-dimensional plane or space by measuring the length of the segment connecting two points. This technique is best for representing the distance between the users and the AP. The collection of RSS data uses the Fingerprinting technique. The RSS data was collected from 20 APs detected using the wifi analyzer application, from the results of the scanning, 709 RSS data were obtained. The RSS value is used as training data. K-Nearest Neighbor (K-NN) uses the Neighborhood Classification as the predictive value of the new test data so that K-NN can classify the closest distance from the new test data to the value of the existing training data. Based on the test results obtained an accuracy rate of 95% with K is 3. Based on the results of research that has been done that using the K-NN method obtained excellent results, with the highest accuracy rate of 95% with a minimum error value of 5%