Harnessing Topological Descriptors: A Comparative Analysis of Artificial Neural Networks and Random Forest for Predicting Anti-Alzheimer Drug Properties

Abstract

The optimization of drug discovery, particularly for neurodegenerative diseases such as Alzheimer’s, remains a critical focus in pharmaceutical research. This study develops a quantitative structure property relationship model to predict the properties of anti-Alzheimer drug compounds using degree-based topological indices as molecular descriptors. Two machine learning algorithms random forest and artificial neural networks are applied to construct and validate predictive models. The performance of each model is evaluated using statistical metrics, including mean absolute error, mean squared error, root mean squared error and the coefficient of determination. Results show that the artificial neural network model outperforms the random forest model, yielding lower prediction errors and higher accuracy. This methodology demonstrates the effectiveness of combining topological indices with computational models to enhance the prediction of drug properties and support the design of more effective therapeutic agents.