Machine Learning-Driven Density Prediction for Nanomaterials

This paper provides a machine learning method that uses band gap and chemical composition data from the Materials Project database to predict the density of nanomaterials. We developed an improved Random Forest Regressor and compared it against several baseline models, including Linear regression, to demonstrate the superior performance of our approach Using a rigorous preprocessing procedure, we combined elemental characteristics extracted from chemical formulae with band gap data. To improve the random forest hyperparameters and boost the predictive power of the model, we employed grid search cross-validation. Key components that have the biggest effects on nanomaterial density were identified via feature importance analysis. Insights into structure-property correlations in nanomaterials were gained by examining the link between density and band gap. Because machine learning allows for quick density estimates, this work shows how it can speed up the discovery and design of nanomaterials. By enabling high-throughput screening of nanomaterials and directing experimental efforts in materials synthesis and characterization, the created model can be a useful tool for nanotechnology researchers and engineers.