PSI - Issue 68

Oleh Yasniy et al. / Procedia Structural Integrity 68 (2025) 132–138 O. Yasniy et al. / Structural Integrity Procedia 00 (2025) 000–000

134

3

3. Results and discussion As a result of applying machine learning methods such as boosted trees, random forests, SVM, KNN, and ANN, models were obtained to predict the properties of SMA for each method. The input data of the models were the stress values s (MPa) and the number of load-unload cycles N , and the output parameter was the strain e (%). Fig. 1 shows the relationship between the experimental values of material strain ! ! " $ #%! $ and the predicted values &!$ " % $ ! '. for boosted trees. a b

Fig. 1. The predicted versus true (experimental) values of the strain, obtained using the boosted trees method for the period of loading (a) and unloading (b) of the SMA material. Fig. 2 shows the relationship between the experimental values of material strain ! ! " $ #%! $ and the predicted values &!$ " % $ ! '. for random forests. a b

Fig. 2. The predicted versus true (experimental) values of the strain, obtained using the random forests method for the period of loading (a) and unloading (b) of the SMA material. Fig. 3 shows the relationship between the experimental values of material strain ! ! " $ #%! $ and the predicted values &!$ " % $ ! '. for SVM.