LAUSR

Chinese mitten crab (Eriocheir sinensis) needs sensory evaluation for grading. This study compared data mining (DM) and sensory panel evaluation (SPE), using data visualization (DV) and quantitative descriptive analysis (QDA), respectively. Results showed that Yangcheng Lake Crab (YLC) was the most welcomed for "umami" and "sweet" according to DV; and QDA (7-scale) showed similar results of the highest "aroma-sweet" (Average Score 4.5) and "taste-umami" (Average Score 4.6) in YLC. The difference was that, DV was fast based on big data (1.4 million words); while QDA quantified detailed attributes (principle components > 85.3% averagely) based on highly-trained sensory panel of good distinguishing- and repeating- ability that F value showed 76.4% of all attributes > 5% for panelist averagely, and mean square error < 0.500 except one panelist. In conclusion, DM was quick but qualitative; while SPE was laborious but informative.