Classification of Urease Activity in Full-Fat Soybean Production by Extrusion Using Machine Learning Algorithms

Soybean is an important food source that frequently preferred in animal feeds with its high protein value. However, soybeans contain many bioactive compounds are antinutritional and/or poisonous. Urease one of the most these. Processes such as extrusion used to reduce these components' effect. Here, factors steam pressure and temperature affect cooking level product. In case undercooked soybeans, components harm health preserve their effect, while nutritional value decreases overcooking. The urease test has been for years evaluate soybean. according color change on product a result test, evaluated by expert. This process mostly done manually dependent expert judgment. this study, machine learning-based approach proposed images results. Accordingly, samples were taken from extruder during processing full-fat A data set consisting over-cooked, well-cooked sample was prepared performing test. binary classification cooked three classes carried out four different learning models set. way, it aimed both automate minimize problems may arise errors. Classification achievements 96.57% 90.29% achieved, respectively, two class tests CNN-LSTM model 10-fold cross-validation tests.