ABSTRACT
A robust near-infrared (NIR) method is reported for the determination of moisture and ribavirin in effervescent granules. Several key factors were used for model construction to obtain robust and universal models, including the number of batches, operators, days, temperature, and data acquisition. The moisture model covered seven representative moisture concentrations: 0.5, 1.0, 1.5, 2.0, 4.0, 6.0, and 8.0%. In this model, the ribavirin concentration varied from 80 to 120%. The NIR models were developed using algorithms for partial least squares and synergy interval partial least squares regression. The results of the algorithms were all suitable based on traditional evaluation standards. However, the accuracy of the synergy interval partial least squares regression algorithm for the moisture and ribavirin concentrations yielded particularly satisfactory results. The model was validated and its reliability evaluated with respect to the measurement of new pilot batches containing 90 and 110% of the active concentration and of individual industry batches. These findings showed that the NIR method is robust rapid, and nondestructive for the determination of moisture and ribavirin in effervescent granules.