Intelligent Multisensor System For Analytical Control Of Sausages

Abstract

The new technique of intelligent analysis of chemical aroma patterns of boiled sausages obtained by the electronic nose for authentication and microbiological safety assessment is developed. The informativeness of features extracted from steady-state responses of the multisensor system and robustness of chemometric algorithms for solving the objectives of qualitative and quantitative analysis of sausage volatile compounds are investigated. The classification model was built using maximum response values as input vectors of an optimized probabilistic neural network, which allows obtaining a 100 % accuracy of different sample grades identification and detection samples adulterated with soy protein. The method of partial least squares regression and area values as features were used for regression modelling and prediction of QMAFAnM with a relative error less than 12 % for a microbiological safety assessment of previously identified sausages. The use of the robust analytical technique to assess authentication, adulteration, total bacterial count for one measurement using the electronic nose in combination with machine learning algorithms will allow to significantly reduce the measurement time and the cost of analysis, and avoid subjective estimation of the results.