Improvement of hydrological network model using object-based classification based from InfoGain feature selection

This study aims to cover the gap between the geometry and topology in drainage network classification by creating a model produced using a weighted flow function. The proposed methodology focused on building a model from Object-Based Image Analysis (OBIA) based from the sophisticated parameter to be used as the weight input to morphological analysis. This study performed segmentation evaluation using Area Fit Index. Further, to improve classification performance, this study performed feature selection using InfoGain based from the stratified random sampling in python.