Classification of Lower Back Pain Using K-Nearest Neighbor Algorithm

Abstract

Back pain is the second most popular disease after the common cold. More than half of the world's human population is affected by lower back pain in their activities. Lower back pain can be categorized into two categories namely normal and abnormal lower back pain. Lower back pain has symptom similarity then it is difficult to determine whether or not the pain is felt without direct examination. It takes the right diagnosis, then treatment of the disease can be done immediately to prevent the worsening impact. In the previous study, using a classification algorithm such as Support Vector Machine, they developed methods to study the effect of the brain structures that occur in lower back pain patient. This study classified specifically normal or abnormal lower back pain suffered based on 12 Range of Motion (ROM) attributes using the K-Nearest Neighbor algorithm as the proposed method. In this study, we used replace missing value and selected attribute methods to preprocessing the data, and in classifying lower back pain, machine learning algorithms such as KNN, Logistic Regression, Naïve Bayes, Random Forest, Decision Tree, and CART were used with the result of accuracy level compared to previous research increased by 15.96%.