FULL UNIAXIAL COMPRESSION CURVES FOR PREDICTING SENSORY TEXTURE QUALITY OF COOKED POTATOES

The prediction of six different sensory texture attributes of twenty-seven potato samples based on uniaxial compression curve features and full uniaxial compressions curves is reported. The data set comprised five varieties, sorted in dry matter bins, and sampled at two storage times. The predictions of most of the sensory texture attributes from full compression curves performed better than just using curve features. Furthermore, force-deformation curves are shown to give slightly better predictions than stress-strain curves. For most of the texture attributes better predictions were obtained using uniaxial compression data from raw potato samples as compared to cooked samples. This study is meant to open up a debate on interpreting the information in uniaxial compression curves in relation to sensory texture quality. Multivariate analysis can handle this and can further give insight about where in the curve the information relevant for the predictions of the sensory attributes is found.