Boiling is an effective heat transfer mechanism, which is a very common heat transfer phenomenon in a wide range of fields, such as metallurgy, nuclear reactor, new energy, electronics, and so on. The bubble departure frequency is one of the critical parameters determining how fast vapor bubbles can carry the heat out from the heating surface. However, because of the complex behaviors in the phase change process, there are few reliable approaches to predict bubble departure frequency. In this study, a machine learning-based approach is proposed to predict the bubble departure frequency with the help of a consolidated dataset. The consolidated dataset of bubble departure frequency, including four kinds of working fluids in subcooling flow boiling, is formed to demonstrate the efficiency and ease of building and deploying models viz. linear regression, KNN, SVM, and ensemble learning models (Decision tree, Random Forest, AdaBoost, Gradient Boosting, XGBoost, and Bagging). As well as the input parameters, including geometric and dimensionless parameters, are also contrasted for a suitable approach. Generally speaking, the optimized XGBoost model performed better than highly reliable generalized correlations, showed the most significant performance across individual datasets and heating configurations. The machine learning-based approach unlocks a reliable tool for bubble departure frequency prediction
