Artificial neural network models between road traffic noise and urban form indicators in different cities

The relationship between road-traffic noise level and urban form indiciators was analyzed using artificial neural network method. The urban form indicators and road-traffic noise level dataset of Gwangju metropolitan city was divided into training dataset(67%) and test dataset(33%). And decay parameter was changed 0 to e ⁷ by exponential scale, to develop accurate model. 5-fold cross validation was used to validate prediction error. Finally artificial neural network model developed using the data of Gwangju was applied to test dataset of Gwangju and whole data of Cheongju city. Correlation coefficient between noise level from a noise map and artificial neural network model of Gwangju was 0.71 and coefficient of determination was 0.5. And the result of applying artificial neural network model to data of Cheongju were 0.67 and 0.45.