HERBAL DETECTION BASED ON AROMA USING SUPPORT VECTOR MACHINE
Maimunah(1*), Sunarni(1), Seta Samsiana(2)
(1)Universitas Muhammadiyah Magelang
(2) Universitas Islam 45
Email: maimunah[at]ummgl.ac.id
Abstract
The quality of herbal ingredients can be well recognized by the public by inhaling the aroma of herbal plants directly. By knowing the aroma, it can be known the types of herbal ingredients. Based on this, it can be detected according to the aroma of herbs. The herbal ingredients used in this study were temulawak, temumangga and temuireng. Data retrieval of herbal ingredients is done by using an electronic nose. Samples in the form of scents that are captured by the electronic nose system will be processed to form a pattern of changes in the voltage of each sensor used. The electronic nose system used consists of raspberries and sensors TGS2600 and MQ5 which produce data output in the form of voltage data. The output is then classified using support vector machine. The results showed that the support vector machine can classify types of herbal ingredients based on their aroma with an accuracy of 90%.
Keywords: Detection; herbal; e-nose; svm
Topic: Information Engineering