Fast honey classification using infrared spectrum and machine learning

Abstract

Honey has been one previous natural food in human history. However, as the supply cannot satisfy the market demand, many incidents of adulterated and fraudulent honey have been reported. In Taiwan, some common adulterated honey and fraudulent honey incidents include (1) mixing honey with fructose, (2) importing cheap honey abroad but labeling them as domestic honey, and (3) labeling cheaper honey (for example, nectar and lychee honey) as high-price honey (for example, longan honey). It is very difficult for consumers to tell the genuineness of the labeling of honey. To protect consumers and honest honey producers, we aim at exploring and developing an efficient and convenient technology that can effectively classify honey. We analyze the infrared spectra of honey samples and apply machine learning technologies to classify honey. The experimental results confirm that this technology can effectively distinguish several main honey types in Taiwan. This technology has the advantages of non-destruction, immediacy, and low manpower. It can serve as an effective tool to fast screen honey products.

Keywords

• honey
• infra-red
• spectrum
• machine learning
• adulteration
• fraud