Edible Oil Adulteration Detection using CatBoost Regression Model

Edible oil adulteration is a serious problem in these days. Oil is the main source of food. Mixing low quality food substance with the edible oil to increase the quantity is the common activity on adulterations. The primary source of edible oil extracted from ground nuts, mustard seeds, sunflower, coconuts, soybeans, corn etc. oil is the main factor that add fat to the body. Adulteration of low quality substances leads to adding up of unwanted fat to the body. The comprehensive study of various techniques used for analysis of edible oil adulteration is discussed here. The proposed system is focused on creating a lightweight architecture to analyze the edible oil dataset for detection and classification accurately. The proposed approach utilized novel adaptive catboost regression model (ACB) for unique statistical measure based classification process. The current approach uses the Kaggle edible oil dataset and, while taking into account the special attributes Free fatty acid(FFA), Peroxide value(PV), and P-anisidine value(AV), among others, obtained training accuracy of 80% and testing accuracy of 98%. Further the results are compared with existing state of art approaches.