Klasifikasi Tingkat Kematangan Buah Kakao Berdasarkan Fitur Warna Menggunakan Algoritma K-Nearest Neighbor

  • Izha Mahendra Universitas Multi Data Palembang
  • Nur Rachmat Universitas Multi Data Palembang

Abstract

So far, cocoa farmers choose the quality of the maturity level of cocoa pods manually or
make selections based on estimates from these farmers, so that the manual method is very prone
to errors in sorting the quality of cocoa pod maturity with various human factors, such as fatigue
and doubt. Based on these problems, this study developed an application for classification of
cocoa pods using Hue, Saturation, Value (HSV) color extraction with the classification method
using K-Nearest Neighbor (KNN) and applying the evaluation results method using the Euclidean
Distance, so that in choosing the level of maturity Cocoa pods have the same standard and a
higher level of accuracy with digital processing. Therefore this research was conducted. The
process of classification of ripeness into 4 classes, namely: rotten, ripe, unripe and half ripe. With
the KNN classification method, and the dataset used is 80 databases, as well as 40 testing data.
The highest value is at k=1 with 90% accuracy, 90% precision, and 90% recall. The tool used to
develop the system is matlab.

Published
2023-10-10
How to Cite
Mahendra, I., & Rachmat, N. (2023, October 10). Klasifikasi Tingkat Kematangan Buah Kakao Berdasarkan Fitur Warna Menggunakan Algoritma K-Nearest Neighbor. Jurnal Algoritme, 4(1), 31-42. https://doi.org/https://doi.org/10.35957/algoritme.v4i1.5485