keywords: Machine learning, complex systems, Energies, Model, computational techniques
Stemming from the relevance of binding energy in understanding the formation, reactivity, and stability of molecular species, particularly in astrochemical environments, this current study leverages an ensemble model integrating six machine learning algorithms: Bagging, Linear Regression, Random Forest, Gradient Boosting, Bayesian Ridge, and Ridge Regression to predict the binding energies of astrochemically relevant molecules. Gradient Boosting demonstrated superior performance in capturing predictive accuracy and variance among individual models. The ensemble model surpassed the predictive power of single algorithms, offering a robust framework for complex chemical systems. The correlation between predicted binding energies and desorption temperatures provides insight into molecule-surface interaction strengths. The ensemble approach illustrates the potential of machine learning techniques in solving intricate astrochemical problems. The ensemble methods effectively capture complex relationships within the molecular data, leading to more accurate and reliable predictions. The results obtained here can be applied in astrochemistry and material sciences and further stress the relevance of machine learning in predictive modeling in Chemistry and other related fields.