Stochastic gradient descent for knee osteoarthritis classification

Ruhul Selsi, Zuherman Rustam, Sri Hartini, Glori Saragih, Muhammad Ariq Yusaputra Bahar

Abstract


Knee osteoarthritis is one of the most viral disease that spreads around the world and commonly affects the elder within the bones, the cartilage, and the synovium in the knee joint. Based on the article from Arthritis Foundation, more than 27 million people in the U.S. have osteoarthritis, with the knee being one of the most commonly affected areas. Women are more likely to have osteoarthritis than men. According to some studies, knee osteoarthritis is caused by overweight, even age (older than 45), and also genes. The symptoms are when the knee is painful and stiff at times. If a person is diagnosed with knee osteoarthritis, the doctors will try to figure out the stages to get the right treatment. In this era, machine learning will help a lot on the process of classification stages. The author proposed Stochastic Gradient Descent method to classify knee osteoarthritis disease. This method works as optimizer whose objective is finding the right parameter values as it provides minium value of cost function. One of the advantages of this method is: it does not use memory as much as the old Gradient Descent, then it can converge faster and the result of this method is around 85.22% accuracy, so that this method can be a capable analysis tools.

Keywords


Classification; Gradient descent; Knee osteoarthritis; Machine learning; Stochastic gradient descent



DOI: http://doi.org/10.11591/ijai.v10.i2.pp%25p

Refbacks

  • There are currently no refbacks.


View IJAI Stats

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.