Robot Classifying of Gas using Support Vector Machine Method

Moch Fachri (Department of Electrical Engineering, Sriwijaya State Polytechnic)
Nyayu Latifah Husni (Department of Electrical Engineering, Sriwijaya State Polytechnic)
Ekawati Prihatini (Department of Electrical Engineering, Sriwijaya State Polytechnic)

VOLT : Jurnal Ilmiah Pendidikan Teknik Elektro
Vol 3, No 1 (2018): April 2018

Fires often happen in the environment of industrial chemical plant caused by harmful gases. To minimize the incident that can be triggered by the gas it takes a tool capable of classifying gases are in the environment industry. The purpose of this research is to know the success in classifying gas on a fire that was triggered by the dangerous gases. We offer solutions in design and build a mobile robot that can classify objects contain hazardous gases by using the method of pattern recognition, the age of the SVM is still relatively young. Nevertheless, the advantages of SVM compared to another method lies in its ability to find the best hyperplane that separates the two class. Based on the results of testing data can classify SVM managed in accordance with the class. The degree of accuracy achieved SVM in classifying reached 86.66 %.