Predicting the Occurrence of Landside at Penang Island, Malaysia, through Artificial Neural Networks Model

The occurrence of recurring landsidesthat plagued Penang Island in 2017 hasattracted much attention across Malaysia since the incident puts people’s life in danger and worsensthe economy of state and country. Based on the harmful effects, the requirement for predicting the occurrence of landslide is deemed crucial for disaster management agencies to promote awareness and prepare for necessary action. Therefore, this study aims to investigate the meteorological factors, and finally to develop a predictive model based on Artificial Neural Networks (ANN). Two meteorological factors, which are daily mean temperature and daily rainfall are considered in this study based on current data from Malaysian Meteorological Department. As a result, only two percent misclassification rate is recorded by the model compared to the actual data,while its prediction performance is better than regression method. Besides, the correlation between the rangeof daily rainfall amount withcorresponding daily mean temperature for the occurrence of landslide is anovel contribution of this research as little attention had been given by previous studies. The developed ANN model is helpful in predicting the occurrence of landslide inPenang Island and provides beneficial guidance for the disaster management agencies to save properties and people’s lives.