中国油脂
中國油脂
중국유지
CHINA OILS AND FATS
2015年
4期
52-54
,共3页
詹洪磊%宝日玛%戈立娜%赵昆
詹洪磊%寶日瑪%戈立娜%趙昆
첨홍뢰%보일마%과립나%조곤
地沟油%太赫兹%聚类分析%概率神经网络
地溝油%太赫玆%聚類分析%概率神經網絡
지구유%태혁자%취류분석%개솔신경망락
swill-cooked dirty oil%terahertz%cluster analysis%probabilistic neural network
基于普通食用油和地沟油的太赫兹吸收光谱,利用0.16~1.30 THz频域的吸收谱,进行聚类分析,8种地沟油组成一类,食用油自成一类。随机选择了2种地沟油作为验证样品,将其余油样进行聚类,采用概率神经网络法对验证样品进行判定,并成功将2种油判定为地沟油。结果表明,太赫兹时域光谱技术结合统计方法可推广成为鉴别地沟油的快速有效手段。
基于普通食用油和地溝油的太赫玆吸收光譜,利用0.16~1.30 THz頻域的吸收譜,進行聚類分析,8種地溝油組成一類,食用油自成一類。隨機選擇瞭2種地溝油作為驗證樣品,將其餘油樣進行聚類,採用概率神經網絡法對驗證樣品進行判定,併成功將2種油判定為地溝油。結果錶明,太赫玆時域光譜技術結閤統計方法可推廣成為鑒彆地溝油的快速有效手段。
기우보통식용유화지구유적태혁자흡수광보,이용0.16~1.30 THz빈역적흡수보,진행취류분석,8충지구유조성일류,식용유자성일류。수궤선택료2충지구유작위험증양품,장기여유양진행취류,채용개솔신경망락법대험증양품진행판정,병성공장2충유판정위지구유。결과표명,태혁자시역광보기술결합통계방법가추엄성위감별지구유적쾌속유효수단。
Cluster analysis ( CA) was used to analyze edible oil and swill-cooked dirty oil based on the terahertz absorption spectra at 0. 16-1. 30 THz. Eight kinds of swill-cooked dirty oils were clusterd to-gether and edible oil was separated. Two kinds of swill-cooked dirty oils were selected randomly from nine samples as validation samples, and the rest samples were analyzed by CA. Based on the built CA model, probabilistic neural network ( PNN) was adopted to recognize the validation samples, and swill-cooked dirty oils were accurately discerned. It was demonstrated that terahertz technology combined with statistical methods could be extended to become a fast and effective way to discern swill - cooked dirty oil.