ECISION SUPPORT SYSTEMS PRAKIRAAN CUACA HARIAN BERBASIS SEMI-SUPERVISED LEARNING MENGGUNAKAN RECURSIVE K-MEANS DI BANDAR UDARA JUANDA SURABAYA

Abstract

Weather is an air condition in the atmosphere layer somewhere. There are many human activities that
depend on weather conditions. Based on this, weather forecasting is able to solve problems that occur around. One of the
human activities that is very influential on climate change is flight activities at Juanda Airport in Surabaya. Where all
activities that occur in aircraft flight are determined based on weather conditions observing surface air observations
which are then sent to the analysis and weather forecasting section. The process of determining weather forecasts with
manual observation methods often occur errors. Based on this, research is conducted using learning methods based on
weather history data in the hope that it can provide weather forecasts that are more efficient and have a higher level of
accuracy. The method used is semi supervised learning using recursive k-means. Semi supervised learning uses recursive
k-means is a method of development of the k-means clustering method. This study uses five weather variables:
Minimum temperature, maximum temperature, average temperature, humidity and duration of solar radiation. the best
learning process is done with a threshold value of ten which results in an accuracy rate of 100%. And to make it easier
for users to use this system, this research also creates an interface system that can be used by users.