Risk factor analysis of device-related infections: value of re-sampling method on the real-world imbalanced dataset.

The application of re-sampling techniques can generate useful synthetic samples for the classification of imbalanced data and improve the accuracy of predicting efficacy of CIEDI. The peri-operative assessment should be intensified in male and aged patients as well as patients receiving replacement procedures for the risk of CIEDI.