Amanna Dzikrillah Lazuardini
Universitas Nusa Putra

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

Found 1 Documents
Search

Platform Digital Berbasis AI untuk Memprediksi Keasaman Kopi dari Variabel Lingkungan: Bukti dari Kabupaten Bandung Amanna Dzikrillah Lazuardini; Anggun Fergina
Jurnal Sosioteknologi Vol. 25 No. 1 (2026): MARCH 2026
Publisher : Fakultas Seni Rupa dan Desain ITB

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.5614/sostek.itbj.2026.25.1.5

Abstract

Despite the importance of environmental control in specialty coffee production, predictive and data-driven systems that connect environmental conditions with coffee acidity remain limited, particularly in smallholder farming contexts. This study aims to design and implement a digital platform based on the Gradient Boosting algorithm to analyze the relationship between environmental factors (soil, topography, and climate) and coffee acidity (pH) levels in Bandung Regency. Through a machine learning approach, the developed model successfully captures complex non-linear relationships between environmental variables and coffee acidity, achieving a high level of accuracy (R² = 0.95) and reducing the RMSE from 0.077 to 0.040 within five learning iterations. The most influential environmental factor affecting coffee acidity was altitude (0.42), followed by soil pH (0.25) and rainfall (0.18). The predictive model was integrated into a web-based system (KopiAsa), allowing farmers to input environmental data and obtain real-time acidity predictions. This platform functions as a data-driven decision-support tool that enhances analysis efficiency, optimizes farm management, and strengthens the competitiveness of local coffee. The results accelerate the digital transformation of the agricultural sector by supporting precision agriculture and promoting more sustainable coffee production practices in Bandung Regency.