Prediction system of carbon monoxide toxic gas monitoring using double moving average method

Abstract

Abstract Health information at this time is very essential to support the continuity improvement of human life. One of the causes of problems to health in humans is the air pollution. It is necessary to have a monitoring system to provide information on air pollution. In this research, a monitoring system for air pollutant gas, specifically Carbon Monoxide (CO) was developed and calibrated with a prediction system using the double moving average method. The prototype for monitoring CO gas used MQ7 sensor, connected with an Arduino microcontroller and GPRS module. Data readings and prediction results were then displayed in real time connected to the web server. The result of developed system were carried out in the KIM area Port Belawan I, by measuring using CO gas prototypes and sending data in real time to the web server in the internal time of 1 hour 13 minutes. We collected 44 log data from the measurement results of data received from CO gas prototypes. The time range of predictive data with real-time gas CO data is 4 minutes on average. The prediction results using the double moving average method have a maximum mean average percentage error MAPE of 3.39% and a minimum MAPE of 0.74%. While the maximum mean square error MSE is 5.84 and minimum MSE of 0.27.