Autentikasi Kopi Single Origin Indonesia Menggunakan Spektroskopi Fluoresens dan Analisis Kemometrik

Abdullah, Fahd (2025) Autentikasi Kopi Single Origin Indonesia Menggunakan Spektroskopi Fluoresens dan Analisis Kemometrik. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009211134-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Sebagai produsen kopi terbesar keempat di dunia, Indonesia memiliki potensi besar dalam mengembangkan kopi single-origin dari berbagai daerah sehingga autentikasi geografis menjadi perlu untuk memastikan keaslian produk di pasar global. Penelitian ini mengeksplorasi metode spektroskopi fluoresens yang dikombinasikan dengan Principal Component Analysis (PCA) dan Linear Discriminant Analysis (LDA). Sebanyak 19 kelas sampel dikelompokkan menjadi dua grup berdasarkan waktu pengukuran dan distributor yang berbeda. Data spektral diproses menggunakan SNV dan turunan pertama, serta diperluas melalui augmentasi spektral. Analisis PCA menunjukkan pola pemisahan yang konsisten, dengan fitur dominan pada rentang 420-460 nm yang diasosiasikan dengan senyawa fenolik dan alkaloid (seperti asam kafeat, CGA, dan kafein), serta kontribusi tambahan dari rentang 460-530 nm (flavonoid) dan di atas 530 nm (lipid). Analisis Area Under Curve (AUC) pada spektrum ternormalisasi memperkuat hasil ini, dengan perbedaan rasio AUC antar kelas sampel menunjukkan bahwa rentang 420-460 nm merupakan fitur spektral paling diskriminatif untuk membedakan origin. Model klasifikasi LDA menunjukkan performa baik, dengan akurasi uji sebesar 93,0 % (Spektral orde 1) pada Grup 1 dan 92,2 % (Spektral orde 0) pada Grup 2. Namun, performa model tetap bergantung pada kualitas data awal, di mana turunan pertama kurang efektif untuk data dengan intensitas rendah dan noise tinggi. Secara keseluruhan, kombinasi antara spektroskopi fluoresens, preprocessing spektral, augmentasi data, serta analisis kemometrik (PCA dan LDA) terbukti efektif untuk autentikasi geografis kopi single-origin Indonesia.
==================================================================================================================================
As the fourth-largest coffee producer in the world, Indonesia holds significant potential in developing single-origin coffee from various regions, making geographical authentication essential to ensure product authenticity in the global market. This study explores the use of fluorescence spectroscopy combined with Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). A total of 19 sample classes were grouped into two sets based on measurement time and different distributors. Spectral data were processed using Standard Normal Variate (SNV) and first derivative, and further expanded through spectral augmentation. PCA revealed consistent separation patterns, with dominant features observed in the 420-460 nm range, associated with phenolic and alkaloid compounds (such as caffeic acid, CGA, and caffeine), along with additional contributions from 460-530 nm (flavonoids) and above 530 nm (lipids). Analysis of the Area Under Curve (AUC) on normalized spectra further supported these findings, showing that the 420-460 nm region was the most discriminative spectral feature for origin differentiation. LDA classification models achieved high performance, with test accuracies of 93.0% (first derivative) for Group 1 and 92.2% (SNV or Zero Order Spectral) for Group 2. However, model performance remained dependent on initial data quality, with the first derivative being less effective on low-intensity, high-noise data. Overall, the combination of fluorescence spectroscopy, spectral preprocessing, data augmentation, and chemometric analysis (PCA and LDA) proved effective for the geographical authentication of Indonesian single-origin coffee.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Augmentasi Data Spektral, Kopi Single-Origin, LDA, PCA, Spektroskopi Fluoresens, Fluorescence Spectroscopy, Single-Origin Coffee, Spectral Augmentation
Subjects:	Q Science > QA Mathematics > QA278.5 Principal components analysis. Factor analysis. Correspondence analysis (Statistics) Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics) Q Science > QC Physics > QC451 Spectroscopy Q Science > QD Chemistry > QD96F56 Fluorescence spectroscopy
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Fahd Abdullah
Date Deposited:	04 Aug 2025 12:00
Last Modified:	04 Aug 2025 12:00
URI:	http://repository.its.ac.id/id/eprint/125788

Actions (login required)

View Item