Ensemble learning prediction method for predicting radiation swelling of nuclear reactor cladding materials

Abstract

The invention provides an ensemble learning prediction method for irradiation swelling of nuclear reactor cladding materials, relating to the technical field of ensemble learning material prediction combining the results of a plurality of weak supervision models. The method adopts a stacked multi-layer heteromorphic regressor model, which is a two-layer structure. The first layer comprises four different base learners, which are an artificial neural network, a support vector machine, a gradient lifting and a random forest, and the first layer adopts 5 fold cross verification training, and thesecond layer is established by XGBoost. The method can reduce the deviation and the variance, improve the generalization ability of the model, and make the prediction result of the material characteristic more accurate.

Description

technical field [0001] The invention relates to the technical field of integrated learning material prediction combining the results of multiple weak supervision models, in particular to an integrated learning prediction method for nuclear reactor cladding material radiation swelling of stacked multi-layer heterogeneous regressors. Background technique [0002] Machine learning (Machine Learning) methods have been gradually applied to material modeling to predict material properties more accurately. The key is to find a mapping model that projects the input space to the output space for a class of problems, and use this learned model to predict actual data. Machine learning methods commonly used in material design and property prediction include artificial neural network (Artificial Neural Network), support vector machine (Support Vector Machine), decision tree (Decision Tree), etc. Technology research, performance research, and post-irradiation performance research provide...

AI Technical Summary

Problems solved by technology

[0005] In the field of irradiation swelling of existing cladding materials, it is difficult to observe the swelling mechanism from the swelling incubation period and the transition period to the linear change period through experiments. It is urgent to provide a new material prediction method to reduce the bias (Bias) and variance (Variance), Improve the generalization ability of the model and make the prediction results of material properties more accurate

Images