Authentication of Indonesian Macroalgae Based on Vis-NIR Spectroscopy Data Combined with Chemometrics

FRYSYE GUMANSALANGI, Dr. Widiastuti Setyaningsih, S.T.P., M.Sc.; Dr. Manikharda, S.T.P., M.Agr.

2022 | Tesis | MAGISTER ILMU DAN TEKNOLOGI PANGAN

Makroalga merupakan bahan baku penting bagi banyak industri, menghasilkan berbagai produk turunan bernilai ekonomi tinggi. Setiap makroalga memiliki komposisi unik yang dapat memberikan informasi fisik dan kimia tertentu yang dapat digunakan sebagai penanda untuk autentikasi. Komposisinya dimungkinkan berbeda tergantung pada berbagai faktor, termasuk asal wilayah geografisnya. Penerapan teknik eksplorasi tanpa pengawasan, yaitu principal component analysis (PCA), hierarchical cluster analysis (HCA), dan teknik pengawasan nonparametrik seperti support vector machine (SVM) dan random forest (RF) ke data spektroskopi Vis-NIR dilakukan untuk standarisasi kualitas makroalga berdasarkan tiga zona regional di Indonesia (Barat, Tengah, Timur). Sebanyak 35 sampel makroalga dari enam pulau di Indonesia dianalisis. Hasil PCA dan HCA menunjukkan kecenderungan sampel terdistribusi dan mengelompok sesuai dengan jenis spesiesnya. Sementara itu, SVM berhasil mengklasifikasikan sampel berdasarkan zona regionalnya, sedangkan jika dikombinasikan dengan 5-Fold Cross-Validation, mencapai akurasi 82%. Algoritma model RF memperoleh akurasi masing-masing 100%, 80%, dan 82% untuk masing-masing set pelatihan, pengujian, dan 5-Fold Cross-Validation.

Macroalgae is an essential raw material for many industries, producing high economic value of various derived products. Each macroalga has a unique composition that might provide specific physical and chemical information that can be used as markers for authentication. Their compositions may differ depending on different factors, including geographical regions. The application of unsupervised exploratory techniques, namely principal component analysis (PCA), hierarchical cluster analysis (HCA), and nonparametric supervised techniques such as support vector machine (SVM) and random forest (RF) to the Vis-NIR spectroscopic data was performed to standardize the quality of macroalgae based on three regional zones in Indonesia (West, Center, East). A total of 35 macroalgae samples from six islands in Indonesia were analyzed. The PCA and HCA results present a tendency for the sample to be distributed and clustered according to the type of their species. Meanwhile, SVM successfully classified samples based on their regional zones, while in combination with 5-Fold Cross-Validation, achieved an accuracy of 82%. The RF model algorithm obtained an accuracy of 100%, 80%, and 82% for the training, test, and 5-Fold Cross-Validation, respectively.

Kata Kunci : seaweed, geographical origin, exploratory study, Support Vector Machine, Random Forest