<![CDATA[As the second-largest national shallot producer, Nganjuk Regency requires significant effort in calculating its land area. The irresponsible Eye Estimate calculation is still used today. Remote sensing offers an alternative using Sentinel-2 medium resolution satellite imagery as a source. Sentinel-2 spectral bands in June and August 2020 were extracted into basic and composite variables, then trained for machine learning modeling. Internal evaluation of reduced variables of performance and Mc Nemar's test showed the Support Vector Machine (SVM) was good in June object, while the Random Forest (RF) In August 2020. External evaluation of the total difference in the land area against the data published by Statistics Indonesia showed the SVM was good for the objects by the smallest total difference, 897.53 and 5382.48 hectares. A high total difference is not expected. So, the selection and development of models, intensive labeling, or variable selection methods can be used for future research.]]>
Copyrights © 2022
