Evaluation of beef flavor attribute based on sensor array in tandem with support vector machines

Beef is an important red meat with abundant nutrition, and flavor is one of the most significant factors influencing beef quality. In this study, sensor array in tandem with support vector machine (SVM) was used to predict beef flavor attribute. Sensor array consisting of 12 ion-selective electrodes and 1 reference electrode was used to collect ion signal from beef samples. Besides, data acquisition card (DAQ card) and electrochemical workstation were applied to convert the ion signal to voltage signal, respectively. For the data analysis, SVM technique was used to build forecasting models for evaluating M. longissimus dorsi (LDs) flavor by combined voltage signals from DAQ card. Besides, data from electrochemical workstation were analyzed with SVM as well, which was applied to verify accuracy of data from DAQ card. The SVM with radial basis kernel function showed a better result with accuracy of 90% using data from DAQ card, and the accuracy of electrochemical workstation reached 90%. Therefore, it was possible to confirm that the integration of sensor array and SVM analysis provides an effective way for evaluating the flavor attribute of LDs. What’s more, the results of this research also indicate that DAQ card can replace the electrochemical workstation when converting the ion signal to voltage signal with a better performance.