TY - JOUR
T1 - Critical value determination on similarity of fingerprints
AU - Fang, Kai-Tai
AU - Liang, Yi-Zeng
AU - Yin, Xiao-lin
AU - Chan, Kelvin
AU - Lu, Guang-Hua
N1 - Funding Information:
The work was partially supported by the Hong Kong Baptist University Grant FRG/02-03/II-62 and the third author was supported by the Hong Kong Baptist University as a University Fellow.
Publisher copyright:
© 2005 Elsevier B.V. All rights reserved.
PY - 2006/5/26
Y1 - 2006/5/26
N2 - A chemical or DNA fingerprint can be treated as a multi-dimensional vector. Correlation coefficient between two fingerprints which is easy to understand and simple to compute has been widely used to assess the similarity of fingerprints in chemical and Chinese Medicine. In the process of fingerprint examination, we are often confronted with such question on how to assess whether a new tested fingerprint is qualified or not. Usually, the mean or median fingerprint of a group of representative fingerprints measured accurately is regarded as a standard fingerprint, then correlation coefficient between the standard fingerprint and a new tested fingerprint is calculated. The critical value for the correlation coefficient is for assessment of a new tested fingerprint whether it is qualified or not. In this study, a bootstrap method is used to estimate the sampling distribution of the correlation coefficient between the standard fingerprint and a new tested fingerprint under the hypothesis that the new tested fingerprint belongs to the same group with the fingerprints used to construct standard fingerprint. Furthermore, using the simulated distribution we obtain the corresponding critical value as the criteria for fingerprint examination. This method is illustrated using Chinese Angelica (CA) fingerprint data.
AB - A chemical or DNA fingerprint can be treated as a multi-dimensional vector. Correlation coefficient between two fingerprints which is easy to understand and simple to compute has been widely used to assess the similarity of fingerprints in chemical and Chinese Medicine. In the process of fingerprint examination, we are often confronted with such question on how to assess whether a new tested fingerprint is qualified or not. Usually, the mean or median fingerprint of a group of representative fingerprints measured accurately is regarded as a standard fingerprint, then correlation coefficient between the standard fingerprint and a new tested fingerprint is calculated. The critical value for the correlation coefficient is for assessment of a new tested fingerprint whether it is qualified or not. In this study, a bootstrap method is used to estimate the sampling distribution of the correlation coefficient between the standard fingerprint and a new tested fingerprint under the hypothesis that the new tested fingerprint belongs to the same group with the fingerprints used to construct standard fingerprint. Furthermore, using the simulated distribution we obtain the corresponding critical value as the criteria for fingerprint examination. This method is illustrated using Chinese Angelica (CA) fingerprint data.
KW - Fingerprint
KW - HPLC
KW - Pearson correlation coefficient
KW - Statistical bootstrap
UR - http://www.scopus.com/inward/record.url?scp=33646096959&partnerID=8YFLogxK
U2 - 10.1016/j.chemolab.2005.06.018
DO - 10.1016/j.chemolab.2005.06.018
M3 - Journal article
AN - SCOPUS:33646096959
SN - 0169-7439
VL - 82
SP - 236
EP - 240
JO - Chemometrics and Intelligent Laboratory Systems
JF - Chemometrics and Intelligent Laboratory Systems
IS - 1-2
ER -