%A Fitri Riantika 
%T RANCANG BANGUN ALAT IDENTIFIKASI KEMURNIAN AROMA
BUBUK KOPI BLENDING KOPI LUWAK DAN KOPI ARABIKA
DENGAN ELECTRONIC NOSE MENGGUNAKAN METODE JST

BACKPROPAGATION
%X Coffee is one of the important plantation commodities in the economy, serving as
a source of foreign exchange for the country. With the increasing coffee
consumption in society, particularly due to lifestyle changes and technology, the
quality of coffee is gaining more attention. Aroma is a crucial attribute in
assessing coffee quality. This study aims to build an architecture of Artificial
Neural Networks (ANN) to identify the purity of the aroma of luwak and arabica
coffee powder, and to integrate it into a purity identification tool based on aroma.
The method used is an e-nose sensor with pattern recognition through ANN. The
samples analyzed include 100% luwak coffee powder, 100% arabica, and various
mixtures (50%-50%, 75%-25%, and 25%-75%). The research results show purity
predictions: the 100% mixture reached 98%, 75% was 75%, 50% was 64%, 25%
was 69%, and the 0% mixture was 98%. These results demonstrate the potential

of ANN in identifying the purity of coffee aromas. This result shows that the e-
nose sensor can detect the purity of coffee powder with an accuracy of up to 98%

within 0 to 20 minutes.
Keywords: coffee, sensor, artificial neural networks
Kopi merupakan salah satu komoditas perkebunan yang penting dalam
perekonomian, berperan sebagai penghasil devisa negara. Dengan meningkatnya
konsumsi kopi di masyarakat, terutama karena gaya hidup dan teknologi, kualitas
kopi semakin diperhatikan. Aroma merupakan salah satu sifat penting dalam
menilai kualitas kopi. Penelitian ini bertujuan membangun arsitektur Jaringan
Saraf Tiruan (JST) untuk mengidentifikasi kemurniaan aroma bubuk kopi luwak
dan arabika, serta mengintegrasikannya ke dalam alat identifikasi kemurnian kopi
berdasarkan aroma. Metode yang digunakan adalah sensor e-nose dengan
pengenalan pola JST. Sampel yang dianalisis meliputi bubuk kopi luwak 100%,
arabika 100%, serta variasi campuran (50%-50%, 75%-25%, dan 25%-75%).
Hasil penelitian menunjukkan prediksi kemurnian: campuran 100% mencapai
98%, 75% sebesar 75%, 50% sebesar 64%, 25% sebesar 69%, dan campuran 0%
sebesar 98%. Hasil ini menunjukkan bahwa sensor e-nose mampu mendeteksi
kemurnian bubuk kopi dengan akurasi sampai dengan 98% dalam waktu 0 sampai
dengan 20 menit.

Kata kunci: kopi, sensor, jaringan saraf tiruan
%C UNIVERSITAS LAMPUNG 
%D 2024
%I FAKULTAS PERTANIAN
%L eprints82113