Seleksi Fitur untuk Prediksi Hasil Produksi Agrikultur pada Algoritma K-Nearest Neighbor (KNN)

Delvi Nur Aini (Universitas Islam Negeri Sultan Syarif Kasim Riau, Pekanbaru)
Bella Oktavianti (Universitas Islam Negeri Sultan Syarif Kasim Riau, Pekanbaru)
Muhammad Jalal Husain (Universitas Islam Negeri Sultan Syarif Kasim Riau, Pekanbaru)
Dian Ayu Sabillah (Universitas Islam Negeri Sultan Syarif Kasim Riau, Pekanbaru)
Said Thaufik Rizaldi (Universitas Islam Negeri Sultan Syarif Kasim Riau, Pekanbaru)
Mustakim Mustakim (Universitas Islam Negeri Sultan Syarif Kasim Riau, Pekanbaru)

Jurnal Sistem Komputer dan Informatika (JSON)
Vol 4, No 1 (2022): September 2022

Agriculture is one of the largest economic driving sectors in Indonesia. The Central Statistics Agency (BPS) in 2021 recorded that 37.02% of Indonesia's population worked in the agricultural sector. The problem faced by farmers today is the decline in yields, both in quantity and quality due to unpredictable weather, making it difficult for farmers to choose the types of plants that are suitable for planting. The application of data mining techniques has problems related to the complexity of weather parameters and natural conditions that support agricultural production, so it is very important to do feature selection, namely to form the most relevant features. This study conducted an experiment to determine the effect of implementing the Principal Component Analysis (PCA) selection feature on the performance of the K-Nearest Neighbor (KNN) algorithm which produces the highest accuracy of 99.64% in this study.