Analysis of sediment re-formation factors after ginseng beverage clarification based on XGBoost machine learning algorithm.

The aim of this study was to explore the sediment re-formation factors of ginseng beverages subjected to four clarification ways (11 subgroups) including the ethanol precipitation, enzymatic treatment, clarifier clarification, and Hollow Fiber Column (HFC) methods, based on the Extreme Gradient Boosting (XGBoost) model. The results showed that the clarity of the ginseng beverages was significantly improved by all the clarification treatments, but still formed sediment after storage. HFC method exhibited the highest transmittance, the least sediment, and stronger antioxidant activity in the clarification treatment groups. According to the results of chemical composition analyses and partition coefficients, carbohydrates, saponins, proteins and metal elements were involved in varying degrees in the re-formation of the sediments in ginseng beverage after clarification. Based on the above data, the XGBoost model predicted that protein, Rd, Na, K, and total saponins were the five most important chemical components affecting the sediment re-formation in ginseng beverages.