Tuber ultimate compressive strength (UCS) is a very important sensory property of tuber products. The UCS depends significantly on tuber moisture content and cooking degree. Therefore, it is necessary to monitor the UCS of tuber products during the industrial production. In the present study, Fourier transform mid-infrared spectra and UCS reference values of tuber samples at different drying times were collected. The fitting functions between tuber UCS and drying time were evaluated. Moreover, the fingerprint spectral matrices (1500–900 cm−1) were studied by pre-treatment methods and principal component analysis. Then, the spectral features were linked to the tuber UCS based on multivariate calibration models in terms of principal component regression and support vector machine regression (SVMR). It has been confirmed that tuber UCS values fitted by specific exponential function could be successfully used to develop robust SVMR model.
Bibliographical noteFunding Information:
The authors would like to acknowledge the ERASMUS plus programme of quantitative tools for sustainable food and energy in the food chain (Q-Safe) (Project No: 2014-1-MT01-KA200-000327) supported by European Union, and the UCD-CSC Scholarship Scheme supported by of University College Dublin (UCD) and China Scholarship Council (CSC).
© 2018, © 2018 Taylor & Francis.