Identification of Indoor Harmful Gas to Human Respiratory System using Support Vector Machines

Today, it is among widely studied subjects to determine the air quality of living spaces and take appropriate actions. In this study, indoor air quality was continuously monitored using an array of various gas sensors. The system is modelled to detect any harmful gas among different gasses and give a warning. Support Vector Machines was used in the model, and it was trained by sample gas data that was created using a group of harmful and harmless gasses to human health. Tests on the trained model showed that the proposed model was able to classify gasses as harmful or harmless with a 100% success rate. The time it takes for the system to give a response with such an accuracy rate was significantly short. The speed of the system is also an important contribution of this study. This study showed that Support Vector Machines can be used with high accuracy-high sensitivity and high specificity- in determining gasses that are harmful or harmless to human health while measuring indoor air quality.