Detection of Hazardous gases using Deep Learning on Benchmark Dataset

The leakage of gas emission is a most critical aspect to be considered in House appliances, Industries, Coal Mines, Transportation, etc. Leakage of gas identification at early stage is more important to avoid harmful to human beings and environment. Sensors are used to detect these hazardous gases. Low-cost sensors are less in both sensitivity and reliability to detect the gases from farther distances. In this paper Bi-LSTM model with parameter tuning is applied on publicly available dataset and obtained an accuracy of 93.5%, 92% of sensitivity, and 95% specificity with the public dataset.