Rheumatoid Arthritis (Ra) Prediction Using SPSVM (Subspace Support Vector Machine) Data Mining Techniques

Resumen

Data mining is a technique that is performed on large databases for extracting hidden patterns
by using combinational strategy from statistical analysis, machine learning and database
technology. Further, the medical data mining is an extremely important research field due to
its importance in the development of various applications in flourishing healthcare domain.
Rheumatoid arthritis (RA) is measured as an auto-immune illness that affects the
musculoskeletal system causing inflammatory, systematic, and chronic effects. RA is
generally progressive and diminishes the physical functionality that leads to articular and
fatigue damages. Overall, RA harms bone and joint cartilage, weakening muscle joints, and
destructing joints. In this investigation, medical disorder classification based on RA is done
with Ensemble methods. Real-time RA data has been collected from the Sakthi
Rheumatology center that holds 1000 patient profiles (750-RA affected and 250 non-affected
profiles). This dataset is posed with a classification problem with numerous numerical
features. Three ensemble algorithms, like SVM, Ada-boosting, and proposed SPSVM, were
used in this investigation. These ensemble classifiers use k-NN and Random forest for
baseline measurements of the classifier. Data classification is performed with 10-fold crossvalidation,
in which evaluation is done with performance metrics like Accuracy, Precision,
and ROC. The values of these metrics were compared with baseline algorithms and various
ensemble classifiers. This optimality specifies the efficiency of base classifiers with
ensemble classifiers, which provides substantial improvement.