Predictive Assessment of Mycological State of Bulk-Stored Barley Using B-Splines in Conjunction with Genetic Algorithms

• Main Author: Wawrzyniak, J. (Author)
• Format: Article
• Language: English
• Published: MDPI 2023
• Subjects: application of artificial intelligence; evaluation of fungal contamination; evolutionary algorithm; machine learning; mold development; postharvest grain preservation and storage systems; predictive model
• Online Access: View Fulltext in Publisher; View in Scopus

 Featured Application: A predictive model combining a genetic algorithm and a B-spline curve designed to assess the mycological state of malting barley grain, which could be used as an effective tool supporting modern systems of postharvest grain preservation and storage. Postharvest grain preservation and storage can significantly affect the safety and nutritional value of cereal-based products. Negligence at this stage of the food processing chain can lead to mold development and mycotoxin accumulation, which pose considerable threats to the quality of harvested grain and, thus, to consumer health. Predictive models evaluating the risk associated with fungal activity constitute a promising solution for decision-making modules in advanced preservation management systems. In this study, an attempt was made to combine genetic algorithms and B-spline curves in order to develop a predictive model to assess the mycological state of malting barley grain stored at various temperatures (T = 12–30 °C) and water activity in grain (aw = 0.78–0.96). It was found that the B-spline curves consisting of four second-order polynomials were sufficient to approximate the datasets describing fungal growth in barley ecosystems stored under steady temperature and humidity conditions. Based on the designated structures of B-spline curves, a universal parameterized model covering the entire range of tested conditions was developed. In the model, the coordinates of the control points of B-spline curves were modulated by genetic algorithms using values of storage parameters (aw and T). A statistical assessment of model performance showed its high efficiency (R2 = 0.94, MAE = 0.21, RMSE = 0.28). As the proposed model is based on easily measurable on-line storage parameters, it could be used as an effective tool supporting modern systems of postharvest grain treatment. © 2023 by the author.