Background. Endoscopic surgeries have become an alternative for open procedures whenever possible. For such types of operations, surgeons are required to gain several skills, whose development needs hands-on practice. Accordingly, gaining these skills today is a challenge for surgical education programs. Despite the development of several technology-enhanced training environments, there are still problems to better integrate these technologies into educational programs. For an appropriate integration, it is critical to assess the skill levels and adapt the training content according to the trainees' requirements. In the literature, there exist several methods for assessing these skill levels. However, there are still problems in practice for objective and repetitive assessment. Methods. The present study aims to estimate the skill levels of participants in surgical training programs in an objective manner by collecting experimental data from residents in an endoscopic surgical simulation environment and gathering performance metrics. Results. It is shown that, by comparing the results of a number of classification algorithms for the best accuracy estimation and feature set, the "novice" and "intermediate" skill levels can be estimated with an accuracy of 86%. Conclusions. The outcomes help surgical educators and instructional system designers to better assess the skill levels of the trainees and guide them accordingly. In addition, objective assessments as highlighted in this study can be beneficial when designing technology-enhanced adaptive learning environments.
Keywords: classification; feature selection; skill-based training; surgical education; virtual simulation environment.