Investigating the Effect of Climatic Parameters Predicting the Mortality Rate Due to Cardiovascular and Respiratory Disease with Soft Computing Methods

It can be very important to accurately identify and predict with smart models in disease outbreaks and as a result in mortality statistics. This study was conducted with the aim of comparing the performance of multilayer perceptron (MLP) neural network, radial basis function (RBF) and regression support vector machine (SVR) methods in modeling and predicting the time series of mortality caused by cardiovascular and respiratory diseases based on climatic parameters and pollutants. This study has analyzed the cases of death and climate parameters and pollutants monthly for 8 years (2015-2022) from Shiraz city. The data was divided into two subsets of training (60%) and test (40%). The performance of the models was evaluated using R, RMSE and MAE criteria. According to the results, the MLP model had a better performance in simulating the mortality of cardiovascular and respiratory diseases. Based on the results of the evaluation criteria for the MLP model, in the training phase, the values of R, MAE and RMSE are 0.7556, 18.8465 and 25.0671, respectively. Also, in the test phase, R=0.8234, MAE=16.9137 and RMSE=23.6522 were obtained for the superior MLP model. Inputs of carbon monoxide and relative humidity were maximum in cardiovascular disease mortality and sulfur dioxide and precipitation parameters were most sensitive in respiratory disease mortality. The MLP neural network can be used as an efficient method to detect the behavior of diseases and mortality caused by diseases over time.