Design and Optimization of Rosuvastatin Calcium Orally Fast Disintegrating Tablet Using Artificial Neural Network Based on Multilayer Perceptron Model

The purpose of the current study is to design and optimize Rosuvastatin calcium orally fast disintegrating tablet (OFDT) with the assistance of an Artificial Neural Network (ANN) based Multi-layer Perceptron (MLP) model. Rosuvastatin calcium is commonly employed as a cholesterol-lowering agent. In our previous work established literature raw material data of OFDTs were collected from ۹۲ research articles, which contain compositional and evaluation parameters and the data trained with Machine learning techniques (ML) to evaluate the optimal ingredients which helps further to develop and optimize Rosuvastatin calcium OFDTs using ANN based MLP. Rosuvastatin calcium OFDTs were formulated according to a ۳۲-factorial design (randomized Box-Behnken method), and formulations were compressed using the direct compression method with varying compositions of superdisintegrant (Crospovidone) ۲-۴% binder microcrystalline cellulose (MCC) ۵-۲۰%, Mannitol as a diluent, magnesium stearate (Mg st) as a lubricant, and talc (۱-۳%) as a glidant. The developed formulations were assessed to determine their thickness, hardness, friability, disintegration time, and drug content. ANN was used for optimization, and the MLP model was trained using experimental data until a satisfactory R۲ of ۰.۹۹ and normalized root mean square error (NRMSE) of ۰.۰۲۴ was reached. The compressed tablets (F۱۹) exceeded the desired criteria in terms of thickness (۲.۶mm), hardness (۲.۸ kg), friability (۰.۶%), drug content (۹۹%), and disintegration time (۳۶ sec). The potential use of ANN in pharmaceutical formulation optimization to achieve desired performance characteristics is demonstrated by this work. This study shows the efficacy of ANN with MLP in the development of Rosuvastatin calcium OFDTs.

کلیدواژه ها:

Rosuvastatin calcium ، Design of experiment ، Crospovidone ، Orally Fast ، Disintegrating Tablets ، disintegration time ، Artificial Neural Networks ، Multi-layer Perceptron ، optimization