Prediction of fatigue stress concentration factor using extreme learning machine

In this work, Extreme learning machine (ELM) is utilized to find the nonlinear mapping relationship between the feature representations and its Fatigue stress concentration factor (FSCF) of metal material. With the random hidden parameters, the ELM-based predictor can be fast trained. If two kinds of materials have different FSCF values, they will exhibit differently in the predicting, and vice versa. With this pairwise metric constraint, the resultant FSCF predictor tends to achieve better forecasting performances.187