A new application of WT-ANN method to control the preparation process of metformin hydrochloride tablets by near infrared spectroscopy compared to PLS.

NIR spectroscopy was an effective expeditious and nondestructive technique to analyze various physical and chemical parameters of interest to the pharmaceutical industry. This paper proposed wavelet transform-artificial neural network (WT-ANN) to determinate mean particle size of metformin hydrochloride granulation process, tablet compression force, tablet hardness, tablet active content with two pretreatment method, standard normal variate (SNV) and multiplicative scatter correction (MSC). The proposed WT-ANN method which was applied to control the preparation process of metformin hydrochloride tablets was feasible and demonstrated more accurate as an available non-linear method compared to traditional methods such as partial least squares (PLS).