Author/Authors :
Tahmasi, Jahangir Department of Industrial Engineering and Management Systems - Amirkabir University of Technology - Tehran, Iran , Ahmadi, Abbas Department of Industrial Engineering and Management Systems - Amirkabir University of Technology - Tehran, Iran , Mosallanezhad, Behzad Department of Industrial Engineering and Management Systems - Amirkabir University of Technology - Tehran, Iran
Abstract :
Nowadays, air pollution is one of pressing environmental issues, and it causes different diseases especially cardiac and respiratory ones. The relation between air pollutants (including PM10, PM2.5, CO, SO2, NO2, and O3) and heart-pulmonary diseases (ischemic, angina, pneumonia and chronic obstructive pulmonary disease (COPD)) is studied in Tehran, the most polluted city of Iran. Air quality data related to all pollutants PM10, PM2.5, CO, SO2, NO2, and O3 have been gathered. The relation between the pollutants and number of admitted heart-pulmonary diseases patients is modeled by using radial basis function (RBF), multilayer perceptron (MLP) networks and ANFIS in terms of MSE and correlation coefficient. In each network, pollutants are assumed as inputs and heart-pulmonary diseases is considered as an output of the network. The experiments show that the ANFIS network has more accuracy than MLP and RBF. Moreover, the obtained results in ANFIS network show that correlation coefficient between ischemic and NO2, angina and CO, pneumonia and PM10 and COPD and (PM10 & PM2.5) respectively are 0.7229, 0.7006, 0.81 and (0.7280 & 0.7249).
Keywords :
Air pollution , ischemic heart disease , neural networks , ANFIS