论文部分内容阅读
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9d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。
Objective To evaluate the freshness of fish objectively. Methods Headspace solid-phase microextraction gas chromatography-mass spectrometry (HS-SPME-GC-MS) was used to detect the volatile components of fish at different storage time. The fingerprint of the volatile components in fish was established. The fuzzy C-means clustering (FCM) analysis of characteristic common peaks. In order to verify the result of clustering analysis, a discriminant model of fish freshness neural network was established. Results FCM could divide the fish stored for 9 days into 3 categories (fresh, fresh and decayed) well. The established RBF neural network model can well distinguish the freshness of fish, and the correct classification rate of both training and test sets Reached 100%. Conclusion This method is effective and provides a new method for analyzing and testing the freshness of fish.