Ecobehavior dalam Pengelolaan Sungai dengan Analisis Neural Network

Abstract

Abstract.The using of neural network methods is an alternative analysis to predict community behavior in river protection. Therefore this study aims to describe the factors that must be considered in the use of Neural network analysis in compiling eco behavior models in river management. This study uses a neural network analysis with a backpropagation approach. The data presented in the analysis were 210 respondents representing the farming community managing land along the river. This research is located in Soppeng Regency, South Sulawesi. The stages of the study are 1) Compiling a backpropagation architecture, 2) Conducting a validation of the model and 3) Conducting a test of the model to obtain the accuracy of the model. The results of the study that describe ten factors as input in the analysis of neural networks are not accurately used in the eco behavior model. This is based on an accuracy value of 48.57%. Therefore, as an effort to develop eco behavior models with Neural network analysis, it should be done in-depth by arranging diverse architectures.

Keywords: accuracy, backpropagation, behavior