GC-MS结合模式识别评价鱼新鲜度的研究

Study on evaluating of fish freshness by GC-MS combined with pattern recognition

中文关键词:  气相色谱-质谱法  模式识别  鱼新鲜度  模糊C均值聚类分析

英文关键词:gas chromatography-mass spectrometry  pattern recognition  fish freshness  fuzzy C-means clustering

中文摘要:

      目的 对鱼新鲜度进行客观评价。方法 用顶空固相微萃取气相色谱质谱联用(HS-SPME-GC-MS)检测不同储藏时间下鱼的挥发性成分, 建立鱼肉挥发性物质的特性指纹图谱, 利用模糊C均值聚类法(FCM)分析特征共有峰。为验证聚类分析的结果, 建立鱼新鲜度神经网络判别模型。结果 FCM能较好地将贮藏9 d的鱼可分成3类(新鲜、次新鲜和腐败), 建立的RBF神经网络模型能很好鉴别鱼的新鲜度, 其训练集和测试集的正确分类率都达到100%。结论 此方法效果好, 为分析和检验鱼新鲜度提供了一种新的方法。

英文摘要:

      Objective To objectively evaluate fish freshness. Methods The volatile components of fish stored at different days were detected and analyzed by headspace solid phase microextraction-gas chromato-graphy-mass spectrometry (HS-SPME-GC-MS). According to the characteristic fingerprint of the fish volatile substances, common peaks were analyzed using the fuzzy C-means clustering (FCM). Then neural network discriminant model was built to verify the result of FCM. Results It showed that fish stored at 9 days could be classified into 3 statuses (fresh, second fresh, bad) with the results of FCM. RBF neural network discriminant model could classify data from the fish volatile substances, and the accuracy rate of train set and test set was both 100%。Conclusion The method has good effects and is suitable for the analysis and detection of fish freshness.