| NIM | A1C021056 |
|---|
| Namamhs | SALSA MILA |
|---|
| Judul Artikel | Deteksi Non-Destruktif Adulterasi Kopi Arabika Menggunakan Spektrometer Portabel Dan Model Klasifikasi Machine Learning
|
|---|
| Abstrak (Bhs. Indonesia) | Kopi merupakan salah satu komoditas pertanian yang rentan terhadap adulterasi atau pemalsuan, terutama kopi Arabika. Praktik ini menurunkan mutu dan keaslian produk serta merugikan konsumen. Metode deteksi adulterasi konvensional cenderung subjektif dan kurang efisien, sehingga penelitian ini bertujuan membangun model klasifikasi untuk mendeteksi adulterasi kopi Arabika bubuk secara non-destruktif menggunakan data spektra dari spektrometer portabel. Sampel penelitian terdiri dari kopi Arabika bubuk murni dan kopi yang diberi adulteran berupa jagung, beras, dan kacang hijau. Pengujian non-destruktif dilakukan menggunakan spektrometer portabel berbasis sensor AS7265X untuk memperoleh data spektra, yang selanjutnya digunakan untuk melatih model menggunakan algoritma Support Vector Machine (SVM), Random Forest (RF), dan Linear Discriminant Analysis (LDA). Kinerja model dievaluasi menggunakan confusion matrix serta metrik klasifikasi berupa accuracy, precision, recall, dan F1-score. Pengujian destruktif terhadap parameter fisikokimia (kafein, pH, kadar abu, kadar air, warna, dan total padatan terlarut) dilakukan sebagai data pendukung. Hasil menunjukkan bahwa SVM memberikan akurasi tertinggi sebesar 100% pada data ternormalisasi, yang mampu membedakan kopi murni dari kopi adulterasi sekaligus mengidentifikasi jenis adulteran. Temuan ini selaras dengan hasil uji destruktif yang menunjukkan perubahan karakteristik kopi akibat penambahan adulteran. |
|---|
| Abtrak (Bhs. Inggris) | Coffee is one of the agricultural commodities that is vulnerable to adulteration or counterfeiting, particularly Arabica coffee. Such practices reduce product quality and authenticity, causing losses to consumers. Conventional adulteration detection methods tend to be subjective and less efficient; therefore, this study aims to develop a classification model to detect adulteration in Arabica coffee powder non-destructively using spectral data from a portable spectrometer. The samples consisted of pure Arabica coffee powder and Arabica coffee adulterated with corn, rice, and mung beans. Non-destructive testing was carried out using a portable spectrometer equipped with an AS7265X sensor to obtain spectral data, which was then used to train models with Support Vector Machine (SVM), Random Forest (RF), and Linear Discriminant Analysis (LDA) algorithms. Model performance was evaluated using a confusion matrix and classification metrics including accuracy, precision, recall, and F1-score. Destructive testing of physicochemical parameters (caffeine, pH, ash content, moisture content, color, and total dissolved solids) was conducted as supporting data. The results showed that SVM achieved the highest accuracy of 100% on normalized data, capable of distinguishing pure coffee from adulterated coffee as well as identifying the type of adulterant. These findings are consistent with the destructive test results, which revealed changes in coffee characteristics due to adulterant addition. |
|---|
| Kata kunci | Adulterasi, Kopi Arabika, Spektrometer Portabel, Machine Learning, Non-Destruktif |
|---|
| Pembimbing 1 | Susanto Budi Sulistyo, S.TP., M.Si., Ph.D. |
|---|
| Pembimbing 2 | Dr. Pradeka Brilyan Purwandoko, S.TP. |
|---|
| Pembimbing 3 | |
|---|
| Tahun | 2025 |
|---|
| Jumlah Halaman | 13 |
|---|
| Tgl. Entri | 2025-08-15 16:20:37.010901 |
|---|