KLASIFIKASI TUTUPAN LAHAN PERKOTAAN MENGGUNAKAN NAIVE BAYES BERBASIS FORWARD SELECTION
Abstract
Urban growth as one of the economic symptoms related to theprocess of urbanization and population displacement in a major way fromthe countryside to urban areas has fueled the city's growth issues. Thesedevelopments will bring up a number of problems when faced with thereality of the limited City area. Urban land cover data that has manyattributes with 9 types of target classification using the best attributes ofsearch techniques by applying the forward algorithm selection and naivebayes which merit independently in target and requires only a small amountof training data to determine the required parameter estimation inclassification process where accuracy 87,04% better compared to testingusing random forest algorithm-based forward selection at the level of72,72% accuracy. so it can be inferred with previous studies using randomforest algorithm with 84,42% accuracy. but better with naive bayesalgorithm-based forward selection with increased accuracy percentage levelof 2.62%.
Moraref