Application of K-medoids clustering method for grouping corn plants based on productivity, production, and area of land in East Java

Corn plants has an important role in compliance national and international food needs after rice and wheat. According to Ir. Pending Dadih Permana, Director General of Food Infrastructure and Facilities (PSP) of the Ministry of Agriculture, East Java is one of province that the biggest corn producers in Indonesia. But it still cannot supply the increasing market needs in Indonesia. Therefore, it requires a grouping corn plants based on productivity, production, and area of land in East Java in order to increase level of procurement of corn plants in East Java. In process of grouping, it use K-Medoids algorithm. K-Medoids algorithm as known as Partitioning Around Medoids (PAM), which is variants of K-Means method. This is based on the use of medoids, not from observation of mean that owned by each cluster, with purpose reduce the sensitivity of partitions due to the extreme values in the dataset. This algorithm is an algorithm that can produce data that is not sensitive to outliers because one of object with some big value. These deviations can occur from data distribution. Based on the result of grouping corn plants based on productivity, production, and area of land which has been done, there are 3 clusters with at most 21 members, and at least 8 members. From this research, it can concluded that K-Medoids algorithm can help determine potential of the location of corn producing plants in East Java.