The accuracy prediction of time-series data on the occurrence of pollutants from smoking using machine learning

Recently, indoor activities have increased due to the harmful effect of the fine dust and Covid-19, accordingly, interest in creating a safe indoor environment is emerging. Even though there are various causes of indoor air contamination, and indoor smoking, which is both illegal and violent to health, and it needs to be regulated. In this paper, we use IoT sensors for detecting indoor smoking and predict it in the time series method. IoT sensors feature a large number of devices and a large amount of data. Therefore, the time series method is suitable even though it has limitations in predicting and analyzing data on a short-term basis. The aim of our paper is to predict the accuracy of data through a time series method. As a result, considering the characteristics of cigarettes was found to be the most effective with Long-Short Term Memory (LSTM) Models.