Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Jurnal Teknologi dan Manajemen Informatika
Mohammad Syaifur Rohman
Unknown Affiliation
Computer Science & IT Decision Sciences, Operations Research & Management Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Optimized LightGBM Model for Predicting Total Cup Points of Arabica Coffee using Sensory Cupping Data Arya Rezagama Sudrajat; Ricardus Anggi Pramunendar; Mohammad Syaifur Rohman
Jurnal Teknologi dan Manajemen Informatika Vol. 11 No. 2 (2025): Desember 2025
Publisher : Universitas Merdeka Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26905/jtmi.v11i2.16348

Abstract

Evaluating coffee quality through sensory cupping is essential but inherently subjective, as scoring depends on the consistency and expertise of professional panelists. To improve objectivity, this study applies the Light Gradient Boosting Machine (LightGBM) algorithm to predict the Total Cup Points of Arabica coffee using sensory evaluation data. The dataset, obtained from the Coffee Quality Institute Arabica Reviews (May 2023), contains 1,509 cupping records assessed according to the Specialty Coffee Association (SCA) protocol. Nine sensory attributes aroma, flavor, aftertaste, acidity, body, balance, uniformity, clean cup, and sweetness were used as predictors. The modeling process included data preprocessing, feature selection, hyperparameter tuning using RandomizedSearchCV, and performance evaluation through 5-Fold and 10 Fold Cross-Validation. The tuned LightGBM model achieved an R² of 0.9634 and an RMSE of 0.4673 under the 10-Fold scheme. Comparative analysis showed that LightGBM produced lower prediction error than XGBoost, Random Forest, and Support Vector Regression (SVR) when evaluated under identical default parameter settings. Feature importance indicated that flavor, balance, clean cup, and aftertaste were the most influential contributors to total cup points. The findings provide a reliable computational framework to support more objective, consistent, and efficient coffee cupping assessments
Co-Authors Arya Rezagama Sudrajat Ricardus Anggi Pramunendar