Beef quality assessment using AutoML

Meat is an essential part of human food as it provides nutrients. During meat supply and storage, meat products may lose nutrients and produce harmful compounds for human health. Though the quality of meat can be examined via physical characteristics by human senses such as smell and vision but it is not a reliable method. In this paper, we used AutoML to assess meat quality from data collected using an electronic nose system. Electronic nose system consists of multiple gas sensors which senses the gases extracted from beef during the decay process. AutoML automates the workflow of machine learning and eliminates the need for an expert machine learning engineer. Principle component analysis is used for dimensionality reduction which help to remove noise. Reduced data fed to AutoML resulted in state-of-the-art performance while categorizing beef into fresh, semi-fresh, and spoiled classes.