Performance comparison analysis of different classifier for early detection of knee osteoarthritis

Abstract Knee osteoarthritis (OA) is a very common problem, especially among the elderly. Clinically, magnetic resonance imaging (MRI) or X-ray techniques are used for detecting knee OA. This chapter focuses on using different kernels of a support vector machine (SVM) classifier technique for early detection of knee OA. Knee OA is caused by degeneration and deterioration of the articular cartilage of the knee. The main reason for degeneration of the articular cartilage is weakness of the quadriceps muscles. Degeneration of cartilage is a slow process that takes a lot of time. Therefore if we can find weakness in quadriceps muscles at an early stage, then we can predict knee OA at an early stage as well. With this early detection, precaution and therapies can be prescribed. Surface electromyography (sEMG) sensors can be used to detect the weakness of quadriceps muscles. EMG is used for evaluating and recording the electrical activity of skeletal muscles. It is an instrument that performs EMG analysis and records the data in the form of an electromyogram. In this chapter, we present a comprehensive comparison for a classification task of knee OA by using different kernels of SVM. Our results show that the fine Gaussian kernel of SVM with threefold cross-validation gives the highest accuracy in comparison to other kernels.