Er. Collantes et al., PREPROCESSING OF HPLC TRACE IMPURITY PATTERNS BY WAVELET PACKETS FOR PHARMACEUTICAL FINGERPRINTING USING ARTIFICIAL NEURAL NETWORKS, Analytical chemistry, 69(7), 1997, pp. 1392-1397
The immediate objective of this research program is to evaluate severa
l computer-based classifiers as potential tools for pharmaceutical fin
gerprinting based on analysis of HPLC trace organic impurity patterns,
In the present study, wavelet packets (WPs) are investigated for use
as a preprocessor of the chromatographic data taken from commercial sa
mples of L-tryptophan (LT) to extract input data appropriate for class
ifying the samples according to manufacturer using artificial neural n
etworks (ANNs) and the standard classifiers KNN and SIMCA. Using the H
aar function, WP decompositions for levels L = 0-10 were generated for
the trace impurity patterns of 253 chromatograms corresponding to LT
samples that had been produced by six commercial manufacturers. Input
sets of N = 20, 30, 40, and 50 inputs were constructed, each one consi
sting of the first N/2 WP coefficients and corresponding positions fro
m the overall best level (L = 2). The number of hidden nodes in the AN
Ns was also varied to optimize performance, Optimal ANN performance ba
sed on percent correct classifications of test set data was achieved b
y ANN-30-30-6 (97%) and ANN-20-10-6 (94%), where the integers refer to
the numbers of input, hidden, and output nodes, respectively, This pe
rformance equals or exceeds that obtained previously (Welsh, W. J.; et
al, Anal. Chem, 1996, 68, 3473) using 46 inputs from a so-called Wind
ow preprocessor (93%). KNN performance with 20 inputs (97%) or 30 inpu
ts (90%) from the WP preprocessor also exceeded that obtained from the
Window preprocessor (85%), while SIMCA performance with 20 inputs (86
%) or 30 inputs (82%) from the WP preprocessor was slightly inferior t
o that obtained from the Window preprocessor (87%), These results indi
cate that, at least for the ANN and KNN classifiers considered here, t
he WP preprocessor can yield superior performance and with fewer input
s compared to the Window preprocessor.