IDENTIFICATION OF EUGENOL IN CLOVE OIL BASED ON ARRAY MOS GAS SENSOR USING PRINCIPLE COMPONENT ANALYSIS AND SUPPORT VECTOR MACHINE METHODS
1*Agus Margiwiyatno, Arief Sudarmaji
Recently the identification of Eugenol is mostly and accurately measured by means of MS-GC technique. However, it is difficult to be done by farmers or Small and Medium Enterprises (SMEs). One potential parameter of aromatic substance, particularly clove oil, to be measured is its vapor (gases and volatiles). This paper presents a non-destructive measurement of clove oil in various percentages based on Volatile Measurement System (VMS) using array of gas sensors. We build the VMS that consists of array of gas sensors in a sensor chamber, gas transport unit, acquisition unit, and data recognition units. We utilized 9 Metal Oxide Semiconductor (MOS) gas sensors run in dynamic mode using advanced Temperature Modulation (TM) method. We tested nine frequencies of Temperature Modulation to measure clove oil with three levels, i.e. 66.61%, 70%, and 88.24%. Principle Component Analysis (PCA) and Support Vector Machine (SVM) were used to obtain the best frequency which driven the MOS gas sensors in distinction the three clove oils. We found that the optimal frequency of TM was at frequency 4 Hz 75% with error validation accuracy is 13.56% using PCA and SVM technique.
identification, eugenol clove array, principle, component vecto machine methods