Rain Prediction Clustering in Australia Using the K-Means Algorithm in the WEKA and RStudio Application

Dinar Ajeng Kristiyanti (Ilmu Komputer, IPB University, Indonesia)
Irwansyah Saputra (Ilmu Komputer, IPB University)
Rina Rina (Sistem Informasi, Universitas Nusa Mandiri)

Seminar Nasional Informatika (SEMNASIF)
Vol 1, No 1 (2021): Inovasi Teknologi dan Pengolahan Informasi untuk Mendukung Transformasi Digital

Purpose: The purpose of this study is how to create an ideal cluster in predicting rainfall in Australia based on the percentage of the sum of squares error (SSE) using the K-Means algorithm with WEKA and RStudio applications.Design/methodology/approach: The method or stages applied in predicting rain in Australia are through several stages including Data Collection, Data Pre-processing (including Missing Value handling in it), Data Mining Modeling by applying the K-Means Clustering algorithm using WEKA and RStudio, Validation results with SSE as well as Data Visualization using plots.Findings/result: Based on the results obtained, clusters of 2 with an SSE of 28.0% are ideal clusters for predicting rain in Australia. In the WEKA software, rain clusters are represented by blue nodes, and non-rainy clusters are represented by red nodes. While in the RStudio software, rain clusters are represented by black nodes and non-rainy clusters are represented by red nodes.Originality/value/state of the art: Get the ideal cluster in predicting rainfall in Australia by comparing the results obtained using the WEKA and RStudio applications.