Diagnosis of disease in tomato paste by Bacillus subtilis bacteria, Penicillium fungi and Aspergillus fungi with the help of electronic nose

Maintaining the quality of tomato paste is very important for processing industry. Bacteria, fungal toxins and molds are factors that can cause food contamination and spoilage. The purpose of this research was to investigate the performance of electronic nose in detecting spoilage in tomato paste and also to investigate the changes of some important physicochemical properties due to spoilage in tomato paste. Bacillus subtilis bacteria and Penicillium and Aspergillus fungi were used to spoil tomato paste. Sampling for samples infected with bacteria was carried out in 4-hour intervals for 24 hours, and for samples infected with fungi, it was taken daily for one week. . Quadratic Discretion Analysis (QDA), Artificial Neural Network (ANN), Supppport Vector Regression (SVR), were the methods used to achieve this goal. The QDA method showed a good performance in the classification of bacteria and fungi and was able to detect bacterial growth and spoilage of tomato paste in 6 sampling times with 100% accuracy. Classification accuracy with the help of neural network for bacteria based on sampling time was 86.7% and for samples infected with fungi based on the type of fungus was 90%. The best prediction of the physicochemical properties of the sample infected with bacteria and fungi by the ANN model related to the properties of sediment weight percentage and acidity, respectively, and in the SVR model, it related to the properties of pH and acidity, respectively.