摘要
为了快速检测完整鸡蛋和裂纹鸡蛋的贮藏时间,使用电子鼻对贮藏在温度为27~30℃、相对湿度为50%的恒温箱条件下的完整鸡蛋和裂纹鸡蛋的蛋液提取特征,并使用基于随机子空间的集成学习方法对其进行分类分析。结果表明:集成学习方法能较好地区分不同贮藏天数的鸡蛋,对裂纹鸡蛋样本的分类准确率为84.29%,对完整鸡蛋样本的分类准确率为88.57%。对同一贮藏天数的鸡蛋,使用逐步判别分析区分裂纹鸡蛋样本和完整鸡蛋样本,分类准确率最低为89.3%,最高为100%,说明电子鼻可以很好地识别裂纹鸡蛋和完整鸡蛋在贮藏过程中产生的差异。
In order to quickly detect the storage time of intact eggs and cracked eggs, the features of intact eggs and cracked eggs preserved in an incubator at the temperature of 27 to 30 ℃ and a relative humidity of 50% were extracted by an electronic nose, random subspace ensemble learning method was used for the analysis of those samples′classification. The result indicated that ensemble learning could classify eggs with different storage days very well, the accuracy rate of classification for cracked eggs samples was 84.29% and for intact eggs samples was 88.57%. Using stepwise discriminant analysis to distinguish the samples between cracked eggs and intact eggs with same storage days, the accuracy rate was between 89.3% and 100%. The results indicated that the electronic nose could detect the differences very well, which came up in the storage process of cracked and intact eggs.
作者
姬雪可
郑江霞
杨璐
郑丽敏
JI Xueke1, ZHENG Jiangxia2, YANG Lu1, ZHENG Limin1,3(1.College of Information and Electrical Engineering, China Agricultural University, Beijing 100083; 2.College of Animal Science and Technology, China Agricultural University, Beijing 100193; 3.Beijing Laboratory of Food Quality and Safety, China Agricultural University, Beijing 10008)
出处
《中国家禽》
北大核心
2018年第8期39-42,共4页
China Poultry
基金
公益性行业(农业)科研专项(201303084)
现代农业产业技术体系建设专项资金(CARS-41)
关键词
鸡蛋新鲜度
电子鼻
随机子空间
集成学习
逐步判别分析
egg freshness
electronic nose
random subspace
ensemble learning
stepwise discriminant analysis
