The wear test was performed with TC4 alloy, by changing the normal bite force, sliding frequency and cycles, in artificial saliva. Taking the two results for testing samples and the others for training samples, the radial basis function (RBF) and multilayer perceptron (MLP) neural network model and least mean square (LMS) and K* model were built for predicting wear loss respectively. Then an ensemble learning model was built which integrated all the single models based on the weight determined by mean absolute error. Compared with the testing results, it was obtained that the error for ensemble learning model was between 3 and 4%. Also, its prediction error rate reduced over 50% for the tenth group data, which embodied good stability and high precision on predicting the wear loss for dental restorative material.

The construction of a reasonable and reliable deformation prediction model is of great practical significance for dam safety assessment and risk decision-making. Traditional dam deformation prediction models are susceptible to interference from redundant features, weak generalization ability, and a lack of model interpretation. Based on this, a deformation prediction model that considers the lag effect of environmental quantities is proposed. The model first constructs a new deformation lag influence factor based on the plain HST model through the lag quantization algorithm. Secondly, the attention and memory capacity of the model is improved by introducing a multi-head attention mechanism to the features of the long-time domain deformation influence factor, and finally, the extracted dynamic features are transferred to the ConvLSTM model for learning, training, and prediction. The results of the simulation tests based on the measured deformation data of an active dam show that the introduction of the deformation lag factor not only improves the interpretation of the prediction model for deformation but also makes the prediction of deformation more accurate, and it can improve the evaluation indexes such as RMSE by 50%, the nMAPE by 40%, and R2 by 10% compared with the traditional prediction model. The combined prediction model is more capable of mining the hidden features of the data and has a deeper picture of the overall peak and local extremes of the deformation data, which provides a new way of thinking for the dam deformation prediction model.