Application of fuzzy logic and machine learning methods for food quality control

In this study, the quality of food products is assesses using fuzzy based methods to achieve higher accuracy and to support optimal performance for quality control systems. Three products namely wine, olive oil and bread are selected for this study and efficient fuzzy based machine learning algorithms have been developed. The first phase of study has been modelled to assess the quality of wheat bread using Fuzzy weighted relevance vector machine (RVM). The second phase employs the wine quality through the development of Neuro-Evolution of Augmenting Topologies (NEAT) and Fuzzy optimized neural networks using improved crow search optimization. The third phase of study improvises the assessment of olive oil quality using Artificial Neural Networks (ANN), Support Vector Machine (SVM) and Random Forests (RF). In the first methodology, an efficient bread quality estimation framework using LDA and FWRVM algorithms was presented. In the second methodology, an efficient wine quality estimation model using the NEAT algorithm and hybrid classifier of ICSA-FNN which is a combination of fuzzy logic, neural networks and an improved ICSA algorithm were presented. In the third methodology, an efficient olive oil quality estimation model was presented using the PCA and LDA for feature reduction and SVM, RF and ANN classifiers for prediction. Experiments are performed over the collected datasets for breads, wine and olive oil and the results showed that the proposed methodologies provided better performance than the existing conventional methods.