Nature inspired computation and ensemble neural network to build a robust model for spectral data

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Computational Chemistry

Background:

• Simultaneous determination of active pharmaceutical ingredients (APIs) is crucial for quality control.
• UV spectrophotometry offers a cost-effective analytical technique.
• Developing robust methods for complex mixtures like Itraconazole (ITZ) and Secnidazole (SEZ) remains a challenge.

Purpose of the Study:

• To establish a simultaneous determination method for ITZ and SEZ using UV spectrophotometry.
• To compare the performance of nature-inspired algorithms with ensemble neural networks for spectral data analysis.
• To validate the proposed method for application in pharmaceutical dosage forms.

Main Methods:

• UV spectrophotometry was employed for spectral acquisition.
• Fourteen nature-inspired algorithms coupled with Partial Least Squares (PLS) regression served as baseline models.
• An ensemble neural networks model was developed and optimized for simultaneous quantification.

Main Results:

• The ensemble neural networks model demonstrated excellent performance for ITZ and SEZ determination.
• High coefficients of determination (R²) were achieved for both calibration and test sets for ITZ (0.9995, 0.9954) and SEZ (0.9962, 0.9957).
• Low error metrics (RASE, AAE) indicated the model's reliability and accuracy.

Conclusions:

• The proposed UV spectrophotometry method combined with ensemble neural networks provides an accurate and efficient approach for simultaneous ITZ and SEZ determination.
• This method is suitable for the analysis of active ingredients in pharmaceutical formulations.
• The study highlights the potential of advanced computational methods in pharmaceutical analysis.