Land use classification using space-borne multispectral image using Random Forest algorithm and two resampling methods to set hyperparameters.

Abstract

Machine learning offers huge efficiency on the classification of remotely sensed imagery. The strengths of machine learning include the capacity to handle data of high dimension therefore, this thesis provides an overview of machine learning from an applied perspective. We focus on implementing a RF algorithm through applying it with 2 resampling methods Bootstrap & cross validation to classify the LULC of Landsat 8 multispectral image of Boumerdes province.