KOMBINASI SELF-ORGANIZING MAP NEURAL NETWORK DAN K-NEAREST NEIGHBOR UNTUK KLASIFIKASI OTOMATIS CITRA KELOMPOK BINTIK MATAHARI

Satiabudhi, Gregorius and Adipranata, Rudy and Anwar, Bachtiar and Setiahadi, Bambang and T, ALVIN NATHANIEL (2011) KOMBINASI SELF-ORGANIZING MAP NEURAL NETWORK DAN K-NEAREST NEIGHBOR UNTUK KLASIFIKASI OTOMATIS CITRA KELOMPOK BINTIK MATAHARI. In: Konferensi Nasional Sistem Informatika 2011, 12-11-2011 - 12-11-2011, Denpasar - Indonesia.

PDF Download (143Kb)

Abstract

Observation and forecasting of weather disturbances in the earth and space caused by the appearance of sunspots is one of the tasks of the LAPAN. The sunspots are actually a tremendous explosion (flares) and coronal mass ejection (CME) that may affect electromagnetic field in a location on the earth. In order to support an early warning system that will be made, LAPAN wishes to make a computer application that can automatically classify a group of sunspots. Here we classify groups of sunspots on a sun digital image based on the Modified - Zurich Sunspot Classification System. In this paper we present our experiment results on the sunspots group classification using the combination of Self-Organizing Map Neural Networks and K-Nearest Neighbor. The result of testing is promising because the classification accuracy was 100 for the data that is trained and 89.1 for the data that is not trained.

Actions (login required)