Abstract:

In this study, 205 samples of fried instant noodles from 12 main brands, which occupy 80% to 90% market in mainland China, were collected to explore a new thinking for evaluating the quality of fried instant noodles. At the first stage, K-mean algorithm was applied to analyze sensory evaluation scores of eight sensory evsory evaluation indexes. Secondly, based on the results of clustering analysis at the first stage, a model of quality evaluation with fifteen physical indexes was built by means of decision tree. The final results of this study showed that this model for quality evaluation can achieve a prediction accuracy of 95.61% and successfully demonstrat the relationship between sensory evaluation and machine measurement values. In conclusion, it can provide an objective and quantification method to evaluate the quality of fried instant noodles.

Key words: instant noodle, sensory analysis, data mining, decision tree