“…In the dataset of naval propulsion system described in this paper, the input dimension of the dataset is small, the data volume of this dataset is small, and the output are continuous values. Therefore, the appropriate traditional CBM algorithms are Support Vector Regression (SVR) (Smola and Sch€ olkopf, 2004) and HMM, and the appropriate CBM algorithms based on deep learning are Adaboost (Freund and Schapire, 1997), RandomForest (Breiman, 2001) and Deep Neural Networks (DNN) (Manikandan and Duraivelu, 2021). These algorithms are overall modeled by relevant mathematical principles with the exception of DNN, so it is difficult to adjust the models for different systems without retraining their mathematical principles.…”