Comparison of K-Nearest Neighbor, Support Vector Machine, Random Forest, and C 4.5 Algorithms on Indoor Positioning System
Open Access
- 21 June 2023
- journal article
- Published by Al-Jamiah Research Centre in IJID (International Journal on Informatics for Development)
- Vol. 12 (1) , 302-313
Abstract
Today’s most common Positioning System applied is the Global Positioning System (GPS). Positioning System is considered accurate when outdoors, but it becomes a problem when indoors making it difficult to read the GPS signal. Many academics are actively working on indoor positioning solutions to address GPS's drawbacks. Because WiFi Access Point signals are frequently employed in multiple studies, they are used as research material. This study compares the classification algorithms KNN, SVM, Random Forest, and C 4.5 to see which algorithm provides more accurate calculations. The fingerprinting method was employed in the process of collecting signal strength data in each room of the Terpadu Laboratory Building at UIN Sunan Kalijaga using 30 rooms and a total dataset of 5,977 data. The data is utilized to run experiments to determine the location using various methods. According to the experimental data, the Random Forest algorithm achieves an accuracy rate of 83%, C4.5 81%, and KNN 80%, while the SVM method achieves the lowest accuracy rate of 57%.Keywords
This publication has 28 references indexed in Scilit:
- A Review of Indoor Localization Techniques and Wireless TechnologiesWireless Personal Communications, 2021
- Forecasting Risk of Crop Disease with Anomaly Detection AlgorithmsPhytopathology®, 2021
- Indoor Positioning Using WiFi Fingerprint2018
- The Global Positioning System (GPS)2017
- Random forest in remote sensing: A review of applications and future directionsISPRS Journal of Photogrammetry and Remote Sensing, 2016
- Two-Stage Learning to Rank for Information Retrieval2013
- Data Mining: Practical Machine Learning Tools and Techniques2011
- WIFI-Based Indoor Positioning System2010
- Consumer-Grade Global Positioning System (GPS) Accuracy and ReliabilityJournal of Forestry, 2005
- Efficient C4.5 [classification algorithm]IEEE Transactions on Knowledge and Data Engineering, 2002