Random forest-assisted Raman spectroscopy and rapid detection of sweeteners

Abstract

Food safety is important for human healthand social stability. Of these, excessive using artificial sweeteners in food production will cause irreversible damage to the gastrointestinal tract. It is possible to accurately distinguish sweetener type in the food by using Raman spectroscopy. However, labeling the type generally relies on manual operations, which limits its application in rapid detection scenarios. This study introduced machine learning methods (Random Forest algorithm) into the data classification process of Raman detection that enables fast and efficient sweetener type detection. The results showed that the three sweeteners were identified with an accuracy of 1, 2 and 3. In addition, the detection process for the three sweeteners merely took 5–6 s. Considering the versatility of the methodology, this study provides a novel technological route for the rapid identification of ingredients in the food production.