Comparison of Different Machine and Deep Learning Techniques to Predict Air Quality Index: A Case of Kocaeli Province

Air pollution is increasing day by day with the increase of urbanization and industrialization. Increased air pollution adversely affects our health. Air quality index is used to determine to what extent it affects our health. The air quality index is used to classify the quality of the air. In this study, Kocaeli province, where urbanization and industrialization is high, is selected. The data used in the study has been obtained from the Online Monitoring Center established by the Ministry of Environment and Urbanization to monitor air quality. Air quality index was calculated with the report containing the measurement values of the pollutant gases belonging to Kocaeli, and labeled by separating them into their classes. In order to predict the air quality on the prepared data set, the comparison of different machine and deep learning techniques is conducted. These techniques are k-Nearest Neighbor, Naive Bayes, Logistic Regression, Decision Trees, Random Forest, Support Vector Machines (SVM), Recurrent Neural Networks (RNN), and Long-Short Term Memory (LSTM). According to experimental results, by considering the accuracy and AUC parameter used in the performance evaluation of the classification techniques, the highest accuracy value was observed as 94% with the Decision Trees and the highest AUC value was reported as 98% with the LSTM model.