APPLICATION OF ALGORITHMS OF WAVELET TRANSFORM (WT)-ARTIFICIAL NEURAL NETWORK (ANN) AND WT-ROBUST REGRESSION (RR) TO HPLC DETERMINATION OF INDIGO AND INDIRUBIN

Two algorithms of WT-ANN and WT-RR were modelled and applied to HPLC determination of indigo and indirubin in root and leaves of isatis and in ready-prepared Chinese Medicine called “Balangen” compound granules,making it possible to avoid the tedious separation and to determine these two components directly and simultaneously.The components to be determined were extracted from the sample with chloroform,and the extract was concentrated to a volume of 10 mL quantitatively and the solution was ready for HPLC determination.The Extend C18 (4.6 mm×250 mm,5 μm) chromatographic column and a mobile phase of mixed solution of CH3OH and H2O (95+5) were used in the determination.As shown by the RSD′s obtained,both of the algorithms were satisfactory for the analysis of multi-components in a sample,but WT-ANN (RSD<0.6%) was better than WT-RR (RSD<2.5%) in the application.In the analysis of 3 substantial samples,recoveries in the range of 98.8%-101.5% were obtained.It was shown that the proposed method was feasible to be used as a method of on line analysis in the production-process.