University of Guilan Forest Sustainable Development 1 3 2014 12 21 Prediction of forest roadway using artificial neural network and multiple linear regressions Prediction of forest roadway using artificial neural network and multiple linear regressions 285 296 85 FA Saba Peyrov M. Sc. Graduate, Faculty of Natural Resources, Tarbiat Modares University, Iran Akbar Najafi Associate Professor, Faculty of Natural Resources, Tarbiat Modares University, Iran Seyed Jalil Alavi Assistant Professor, Faculty of Natural Resources, Tarbiat Modares University, Iran Journal Article 2014 02 07 Prediction of roadway is one of the main effective factors on fill and cut slope volume, cost and disturbance in forest road constructions. The objective of this study was to develop models for prediction of forest roadway using artificial neural network (ANN) and multiple linear regression (MLR). For this purpose, 192 cross profiles were measured on the Soordar-Vatashan forest roads. Within each sample, hillside gradient, slope direction, rock share ratio and texture of soil were recorded as the inputs and roadway was recorded as the output. The models were developed by artificial neural network with back propagation learning algorithm, multiple linear regression with stepwise analyses, correlation analyses and independent One-Way ANOVA using MATLAB 7.6.0, R and SPSS 19 software. According to coefficient of determination (R²), multiple correlation coefficient (r) and root mean square error (RMSE) and percent error, the ANN was more successful than regression model in prediction of roadway. (p Prediction of roadway is one of the main effective factors on fill and cut slope volume, cost and disturbance in forest road constructions. The objective of this study was to develop models for prediction of forest roadway using artificial neural network (ANN) and multiple linear regression (MLR). For this purpose, 192 cross profiles were measured on the Soordar-Vatashan forest roads. Within each sample, hillside gradient, slope direction, rock share ratio and texture of soil were recorded as the inputs and roadway was recorded as the output. The models were developed by artificial neural network with back propagation learning algorithm, multiple linear regression with stepwise analyses, correlation analyses and independent One-Way ANOVA using MATLAB 7.6.0, R and SPSS 19 software. According to coefficient of determination (R²), multiple correlation coefficient (r) and root mean square error (RMSE) and percent error, the ANN was more successful than regression model in prediction of roadway. (p https://fsdj.guilan.ac.ir/article_85_f8fa517e9db912d300a5e02bbea97442.pdf