DEVELOPMENT OF MODIFIED RANDOM FOREST ALGORITHM WITH BAYESIAN OPTIMIZATION FOR CLASSIFYING PADDY GROWTH STAGE IN KARAWANG REGENCY, WEST JAVA

• Main Author: Gunawan, Elisabeth
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/75665
• Institution: Institut Teknologi Bandung

Rice (Latin: Oryza Sativa) is an important and strategic agricultural commodity in Indonesia. Indonesia is in the 4th position in the world as the largest rice producer in the world. One of the areas that is the center of rice cultivation in Indonesia, especially on the island of Java, is Karawang Regency. With this fact, Indonesia needs careful planning to maintain food security while obtaining maximum benefits from this commodity. In this Final Project, research was conducted on modifying the Random Forest algorithm in classifying paddy growth stage from Sentinel-1A satellite images based on machine learning. Based on the literature studies that have been conducted, it was found that there are several disadvantages of the Random Forest algorithm, one of which is that it cannot easily interpret the relationship between response variables and predictor variables, which results in the structure of all constituent trees being impractical so that the learning process of this algorithm becomes slow. Therefore, a solution was designed to optimize the Random Forest algorithm in classifying paddy growth stage by performing hyperparameter tuning with Bayesian Optimization.