OPTIMIZATION OF GARLIC CULTIVATION LAND SELECTION USING PCA AND K-MEANS APPROACH IN SPATIAL INTELLIGENT SYSTEM

Abstract

The object of this research is agricultural land in highland areas that have the potential to be planted with garlic. The main problem solved is the difficulty of identifying and selecting optimal land for planting garlic efficiently and objectively, especially in large and geographically complex areas. Special focus is given to data and spatial parameters that affect land welfare. In addition, planting garlic can be a promising business opportunity, especially in areas that have environmental conditions that support its growth. However, garlic production in Indonesia is often unable to meet market demand, resulting in dependence on imports. This is a common problem that can increase the price of garlic on the market. This study aims to increase crop yields and resource utilization efficiency, but also provide adaptive solutions to climate challenges and support national food security by using principal component analysis (PCA) and K-means. Researchers use principal component analysis (PCA) to reduce data dimensions or simplify complex input variables such as altitude, rainfall, temperature, soil type, and others-without losing important information. After that, the K-means clustering algorithm was used to group the areas into several land suitability classes based on the results of the dimension reduction from PCA. The PCA and K-Means methods help in data-based decision making for more efficient agricultural land development. The clustering results can be used by farmers, governments, and agribusiness companies to determine the most suitable locations for planting garlic. The results of the spatial study of garlic cultivation land using PCA and K-means successfully determined spatial land by conducting a classification test with a test accuracy using Inerta of 0.49% and using the Silhouette Score test of 0.89%.