Downloads: 200 | Views: 429
Research Paper | Information Technology | Turkiye | Volume 8 Issue 1, January 2019 | Popularity: 6.8 / 10
Comparison Supervised Learning Algorithms for Spinal-Column Disease
Nuredeen A. A. Matoug, Hajer AL-Jebury, Kemal Akyol, Mohammed Ridha Alsarrar
Abstract: In this study the classification were performed on Vertebral Column study. The dataset that were used in this study was performed and studied on 310 orthopedic patients and their vertebral column parameters were evaluated as 6 features for each patient (pelvic_incidence, pelvic_tilt numeric, lumbar_lordosis_angle, sacral_slope, pelvic_radius, degree_spondylolisthesis). These parameters were obtained from a panoramic image of the spine. The 100 subjects had no spinal vertebral pathology issues (normal). and other 210 subjects were assigned as abnormal (60) patient were suffering from disc dislocation ( disc hernia) and 150 subjects were suffering from spondylolisthesis. So the results for these attributes were assigned two classes normal or abnormal (disc dislocation or spondylolisthesis). In this study classification were compared according to their performance by using different (7) classification algorithms are: ( RF, RF, DT, SVM, Nave Bayes (NB), KNNc, MLP). And each of the classifier were evaluated their performance (Accuracy, sensitivity, specificity, Errors). The highest accuracy and specificity and lowest Errors were recorded about (83.4 % and 100 %, 16 % respectively) in nave bayes classifier. where for the highest sensitivity in Random forest were recorded (81.4 %). In this study were showed that nave bayes classifier were recorded the best classifier performance.
Keywords: Vertebral column dataset, ML classification algorithms Logistic Regression, Random Forest, decision Tree, Support Vector Machine, Nave Bayes, k-nearest neighbors classifier, neural network classifier
Edition: Volume 8 Issue 1, January 2019
Pages: 312 - 315
Please Disable the Pop-Up Blocker of Web Browser
Verification Code will appear in 2 Seconds ... Wait