Perbandingan Algoritma Self Organizing Map dan Fuzzy C-Means dalam clustering hasil produksi ikan PPN Karangantu

Abstract

- Data Fish production data located in the PPN Karangantu in 2017-2021 has a total of 13429.7 tons based on the production results of 58 types of fish over the last 5 years and production data can be compared with the use of SOM and FCM algorithms to obtain the best cluster value. A cluster is one of those groupings that occurs based on the same criteria. The purpose of the comparison of the two algorithms is to determine the type of variety of fish, superior production and known groups of low, medium and high fish species. There are 242 rows that can be a dataset in csv form. To provide convenience in managing data, researchers use Matlab 2017b. Comparison of the two algorithms is based on literacy values and clustering results. Based on the literacy values that occur in both algorithms, SOM has 200 iterations and the FCM algorithm has 88 literacy so that the som algorithm obtains optimal and more effective results for clustering. The results of clustering using som are on clusters low 214, medium 18 and high 10. Meanwhile, in the FCM clustering results, low clusters 4, medium 229 and high 8 were obtained. Based on the results of the study, the SOM algorithm can find out the types of fish varieties, superior production and known types of fish based on the results of clustering in PPN Karangantu.