Novel reliable model by integrating the discrete wavelet transform with fuzzy intelligent systems for the simultaneous spectrophotometric determination of anticancer drug and anti-acquired resistance drug in biological samples.

Simultaneous measurement of drugs used to treat cancer and medications prescribed to overcome resistance to these drugs is important in pharmaceutical formulations and biological samples. In this study, a spectrophotometric method with a hybrid of discrete wavelet transform (DWT), principal component analysis (PCA), fuzzy inference system (FIS), and adaptive Neuro fuzzy inference system (ANFIS) was used to predict concentrations of erlotinib (ERL) and niclosamide (NCM) in binary mixtures and biological samples simultaneously. In DWT, three wavelet families named Daubechies 4 (db4), Symlet 2 (sym2), and Demeyer (DM) were utilized to decompose the absorption of mixtures at five levels, and the resulting matrixes were separately reduced through PCA. The reduced dimensionality outputs were considered as inputs to the FIS and ANFIS models. The performance of DWT-FIS and DWT-ANFIS models were surveyed in terms of the statistical indices, such as coefficient of determination (R), root mean square error (RMSE), average testing error, and mean recovery percentage. The wavelet family of db4 with R of > 0.96 and equal to 1 was selected as the best family for FIS and ANFIS, respectively. In the DWT-FIS model, RMSE values were 0.3081 and 1.113 for ERL and NCM, respectively, while in the DWT-ANFIS model, average testing error values of 5.46 × 10 and 5.16 × 10 were obtained for ERL and NCM, respectively. In the selected wavelet, the mean recovery values for both components were > 97.5 % and > 99.8 % in the DWT-FIS and DWT-ANFIS, respectively. The analysis of the spiked biological samples containing ERL and NCM using the DWT-FIS and DWT-ANFIS indicated relative standard deviation (RSD) < 2.3 % and < 1.9 %, respectively. Compared with the DWT-FIS model, the DWT-ANFIS model revealed a better prediction. The recovery averages from the analysis of urine samples related to the proposed methods and the HPLC technique were compared using the ANOVA method. It can be concluded that proposed chemometric-assisted UV-spectrophotometric methods are efficient, reliable, economical, fast, and easy as alternative methods to chromatographic techniques.