CLASSIFICATION OF THE QUALITY OF HONEY USING THE SPECTROFOTOMETER AND MACHINE LEARNING SYSTEM BASED ON SINGLE BOARD COMPUTER

Abstract

Honey is a natural sweet substance produced by bees with nectar flower ingredients containing various nutrients such as carbohydrates, proteins, amino acids, vitamins, minerals, plant pigments and aromatic components. Honey consists of water (17%), fructose (38.2%), glucose (31.3%), other disaccharides (5%), melezitose (<0.1%), erlose (0.8%) , other oligasakarida (3.6%), minerals (0.2%), amino acids (0.3%). It also contains anti-microbial substances, which can prevent some disease on human. Because it has a high economic value and contains unique substances, honey is often falsified. Honey can be faked in various ways by mixing natural artificial sweetener components or by giving the sugar (sucrose) into the honey, so it is dangerous if given to infants or people suffering from diabetes mellitus. The design of this research used spectrofotometer methode and pattern recognition algorithm (machine learning) system for classifying honey quality. It is based on honey having a chromophore group responsible for absorbance, electronic transition and color giver. We can also knowing the type of honey by using pattern recognition algorithm support vector machine.In this research is used three types of honey to using a spectrophotometer at a wavelength of 500 nm, the absorbance obtained in randu honey type of 0.523-0.654, coffee honey for 0.735-0.824, and rubber honey by 0.947-1.043.