Assessment of the Effect of Training Data Selection on Landslide Susceptibility Mapping

Landslide susceptibility mapping is a component of geo hazard assessment and mitigation planning. This study focuses on analyzing and studying the impact of training data on the accuracy and reliability of landslide susceptibility maps. We analyze different parameters, including size, spatial distribution, and diversity, on the performance of machine learning models employed for landslide susceptibility modeling. We utilize multiple data set comprising geological, topographical, climatic, and many other variables to develop models. Our findings reveal that the selection of training data significantly affects the model's pre dictive capabilities, with implications for both false positive and false negative rates in landslide prediction. This study provides insights into optimizing training data selection strategies for more accurate landslide susceptibility mapping, thereby contributing to geohazard assessment. Landslide occurrences during the rainy seasons pose significant challenges in the Himalayan region and hilly areas of India. Nevertheless, there is a lack of sufficient data and research pertaining to landslides in these vulnerable regions.