KLASIFIKASI KOPI LOKAL MENGGUNAKAN HIDUNG ELEKTRONIK DENGAN ANALISA METODE MULTIVARIAN; LOCAL COFFEE CLASIFICATION USING ELECTRONIC NOSE WITH MULTIVARIATE METHODE ANALYSIS

Ardiansyah, Deny Enggar T, Triyogatama Wahyu Widodo

2015 | Skripsi | FMIPA

Test the quality of foodstuffs is very necessary for the consumer to accept or reject a product. Coffee is a food that has a scent identical or typical. Aroma produced by the aromatic compounds in coffee greatly affect the quality of the coffee. Therefore, a test instrument for testing the aroma of coffee is needed especially during the process of post-harvest quality control. Aroma test instrument used in this study is the electronic nose. Arabica coffee beans (Arabica Coffea) from West Java, Batak and Toraja with postharvest processing of different used as the test samples, each tested with a mass of 10 grams. Data were collected with the sensing process for 120 seconds and flushing for 60 seconds with repetition as much as 5 times. Data processing was carried out in several stages in antranya preprocessing the signal by manipulating the baseline deferential, signal processing with methods of feature extraction mean (average) and multivariate analysis using Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) which is validated by Gas Chromatography Mass Spectrum (GCMS). Value percent cumulative variance of two main components produced in scoreplot PCA to sample the local coffee amounted to 99.2% and prediction accuracy of the discriminant model is at 100%. GCMS analysis showed that there are differences in the amount of aromatic compounds contained by each sample coffee, this is in accordance with the testing of electronic nose with PCA and LDA analysis method that can classify three types of local coffee.

Kata Kunci : Electronic nose; Principal Component Analysis (PCA); Linear Discriminant Analysis (LDA); GCMS; coffee aroma.