A preprocessing method applied in multi-dimension data - NOSC

Orthogonal signal correction (OSC) is a novel spectral preprocessing method, which is based on the orthogonal projection. This pre-processing way removes from the spectral vector or matrix(X) only the part that definitely is unrelated to Y-vector or Y-matrix. Due to the merits of OSC, much interesting is: focus on it, and several modified algorithms have been presented to improve it. In this paper, we present a modified multi-dimension method, named NOSC, which is applied in preprocessing three-dimension data. The three-dimension data is the measurement of a drug mixture by HPLC-DAD, in which there are three components, that are enoxacin, norfloxacin, and ciprofloxacin. Compared with, those only processed by NPLS, the multi-dimension data preprocessed by NOSC have an optimal analytic result, and the covariance sauqre root is 0.33, 0.21 and 0.16 for enoxacin, norfloxacin, and ciprofloxacin respectively. NOSC was recommended for processing multi-dimension data with NPLS, PRFA and GRAFA, etc..