Fast and robust quality assessment of honey using near infrared spectroscopy

Assessing the quality of honey is crucial toupholding industry standards and guaranteeing customersafety. Using important physicochemical characteristics likepH, viscosity, electrical conductivity, glucose, fructose, andpollen analysis, this study uses machine learning approachesto forecast honey purity and pricing. To assess the predictivepower of the RF and Gradient Boosting Regressor models,we put them into practice and compare them. To improvemodel performance, the dataset is preprocessed usingtechniques like feature scaling and categorical encoding.Gradient Boosting outperforms RF in terms of predictionaccuracy, as evidenced by evaluation criteria like R2 score.According to the findings, machine learning can help withhoney quality evaluation and verification in an efficientmanner, offering producers, consumers, and regulatoryagencies useful information.