Research On Noodles Quality Evaluation Method Based On Data Mining

As a traditional main food, a strict, uniform method of noodles' quality evaluation has not established. Aiming at adapting the need of modern examination techniques, food safety and quality inspection, and the target of achieving simple, easy locale examination and integration, objective quality evaluation, it is imperative under the situation that developing a new generation convenient food mechanics performance examination instrument, and establishing an objective evaluation method. Based on the data collection, store and processing of noodles mechanics performance indexes, this dissertation proposes an easy and objective method of noodles mechanics performance evaluation.Based on plenty of researches on the noodles mechanics performance examination status home and abroad, this dissertation analysis and research the possibility of noodles quality evaluation method, and design a experimental method of noodles mechanics performance examinationo We collect plenty of valuable data about the 46 noodles samples through the experiment, which operate a positive effect on noodles mechanics performance evaluation.Based on the original instrument, this dissertation designs the clamper of instrument according to the characters of noodles, which gives a solution of the problem of clamper unstable. This makes the measurement of pull and cut more exactly. This new instrument makes the data indexes examination and data collection come true. For better analysis of the dataset, this dissertation develops a food mechanics performance analysis system based on data mining. This system can achieve the store and analysis of data and it has several characters such as easy use and display the result visually. Users can use this system conveniently to evaluate the mechanics performance of food.This dissertation introduces the data mining technology to the evaluation method of noodles mechanics performance. Clustering analysis is an important method in data mining. K-means algorithm is an important clustering method based on partition. It is simple, fast, high efficiency better retractility. The dataset of noodles samples is compose of each mechanics performance indexes and other attributes. The result of k-means algorithm on this dataset is 5 clusters. Through to evaluation of these clusters, each cluster has the attribute character which can delegate itself. This result is a significative partition of the dataset, which can use to the evaluation of noodles quality. Clustering analysis only can give a qualitative description of noodles quality but only a partition of the noodles. This dissertation compares the clustering result with the PCA and sense evaluation which is considered as a consultant, and than evaluate the S clusters of clustering analysis. The result is the rank about noodles quality and the description of each rank. Through the experiment in this dissertation, the rank of noodles samples is clear, and user can get the intuitionistic information of quality rank, toughness, chewiness, flexibility and taste. And these solve the problem of noodles locale examination.The noodles quality evaluation in this dissertation reflects the food character about the mechanics character in noodles quality. The result is high sensitive and objective. This method can evaluate the noodles quality objectively by the quantitative processing, which avoid the subjective influence by man-made factors in evaluation.