Yücel, EnginAkyuva, Yener2024-09-192024-09-1920231309-4483https://doi.org/10.52142/omujecm.40.3.35https://hdl.handle.net/20.500.12483/14476We aimed to investigate the possibility of determining the relationship between the dominant extremity and cervical disc herniation using a machine learning approach. A total of 561 patients diagnosed with cervical disc herniation were examined for dominant extremity, level and side of cervical disc herniation, and the nature of the herniation (calcified/soft). These patients formed the basis for a two-step machine learning system creation. The first step (included the data of 80% of the patients) focused on determining the type of cervical disc herniation by selecting the top five performing classification models out of 15 different models and tuning the hyperparameters. In the second step, the machine learning system was validated using data from a randomly selected subset of patients (20% of the patients). The study results showed that while most models performed well, the gradient boosting classifier was the most accurate (89.38%) for determining the herniated disc nature. However, for classifying the disc herniation direction, the models did not exhibit strong performance. Thus, machine learning can accurately identify the relationship between cervical disc herniation and dominant extremity with a high degree of accuracy. © 2023 Ondokuz Mayis Universitesi. All rights reserved.eninfo:eu-repo/semantics/closedAccessartificial intelligencedisc degenerationmachine learningspinal diseaseArticle40363363910.52142/omujecm.40.3.352-s2.0-85186901249Q4