Regressive Multi-Scale BiGRU Model for Work Function Prediction of MXene Materials With Oppositional Learning-Based Optimization

MXenes are two-dimensional transition metal carbides and nitrides with variable compositions and complex chemical properties. MXenes are used in various applications such as electronics, energy storage, and catalysis due to their highly tunable electrical, optical, and physical characteristics and composition adaptability. The applicability of MXenes for these applications can be determined by its work function, which is a crucial fundamental characteristic. Many researchers have recently concentrated on machine learning models to predict the work function, but these models still need improvement. So, this research proposes a deep learning (DL) based accurate 2D MXenes work function prediction model called a Regressive Multi-scale Dynamic Adaptive Convolutional BiGRU Weighting (RMDAC-BiGRUW) ensemble model. To improve the effectiveness, the pre-processing stage is included for normalizing and standardizing the input data. Furthermore, the oppositional gazelle transition algorithm (OGTA) technique is employed for deep feature selection. Finally, the proposed DL-based ensemble model, composed of several layers, is utilized for work function prediction. The proposed model is accessed with various existing models in the performance evaluation scenario, including decision tree, Random Forest (RF), K-Nearest Neighbors (KNN), and Convolutional Neural network (CNN). Furthermore, different kinds of analysis are performed in terms of mean absolute error (MAE), mean squared error (MSE), and coefficient of determination (R2). Finally, the experimental outcomes demonstrate that the proposed model is perfectly suitable for precise MXenes work function prediction as the model selection criterion values are better compared to the existing models. In addition, the proposed work function prediction model obtained 0.1875 as MSE, 0.7662 as R2, and 0.1875 as MAE.