ADOLESCENT IDIOPATHIC SCOLIOSIS DETECTION USING A NOVEL MACHINE LEARNING APPROACH

National Scoliosis Foundation (NSF) has found that 2 to 4 percent of the people these days are suffering from scoliosis. Scoliosis is a disease which develops an abnormal curve in the spine region. This disease affects the population comprising all age groups. In US alone about over 7 million people are diagnosed with scoliosis. Adolescent Idiopathic Scoliosis (AIS) is the most common scoliosis type that is typically found in patients belonging to the age groups of 40 years or more due to the problem of aging. Studies have revealed that about 64% to 68% are affected by this disease among the aged population groups. This has become a significant healthcare concern. Demographic, Radiographic and Coronal Cobb Angle measurements were collected to measure the angular difference between the neutral spine and the abnormal spine to estimate the probability of the scoliosis disease in adults, in the affected region. The experimental results were compared with the supervised algorithms kNN and SVM. The performance of Naïve Bayes proved prominent based on the results of prevalence, specificity, sensitivity, error rate and accuracy.