Building regression model to estimate NOXemission pollutant of backhoe equipment

Abstract

NOx emission is one of the gas pollutants resulted from backhoe equipment, non-road diesel equipment usually used in the construction industry. Predicting NOx emission may help the construction industry to assess the environmental effect of the construction project operation. This study aims to build a regression model to estimate NOx emission pollutant of a backhoe equipment using a CRISP-DM approach. The model is built based on 37,529 data points. The predictor variables are engine speed/revolutions per minute, manifold absolute pressure (MAP), engine rated power/HP (watt), intake air temperature, equipment age, and EPA engine tier technology. The best regression model is obtained by setting the split ratio on 0.9. The results shows that MAP has the highest effect, while the EPA engine tier technology type has the lowest effect on NOx emission.