Moisture content prediction model for pharmaceutical granules using machine learning techniques

Abstract

The aim of this study was to develop a prediction model and identifying relative important factors in the evaluation of moisture content of pharmaceutical granules using Artificial Neural Network (ANN) and Support Vector Machine (SVM) techniques for the data sets of Addis Pharmaceutical Factory (APF). Optimal models of ANN and SVM models were developed and compared, utilizing Matlab16.0 as a software tool. The performance of the models is evaluated using a quantitative error metric; mean squared error (MSE), Regression(R) and Confusion Matrix (CM). This study reveals that the ANN model is an optimal model for predicting moisture content of pharmaceutical granules for the datasets of APF. The model of ANN, with MSE of 0.016941 and classification accuracy of 98.7% is built and accepted as optimal model for predicting moisture content of pharmaceutical granules. Temperature, Mixer time, Initial moisture, air flow rate and drying time respectively are the most important factors in determining the moisture content of the granules.