PCA-SVM Classification: Motor Ability of Down Syndrome Based on Education, Economics And Physiotherapy Therapy

The classification of motor abilities of individuals with Down Syndrome is essential to identify more effective developmental patterns. However, research that integrates educational, economic, and physiotherapy factors in the classification model is still limited, especially in the application of machine learning-based methods. The purpose of this study is to classify using PCA-SVM on the motor ability of DS based on education, economics, and physiotherapy therapy. PCA is used to reduce the dimensions of the dataset by extracting the main features that have the greatest variation, thereby increasing the efficiency and accuracy of classification. Meanwhile, SVM with Radial Base Function RBF Kernel is applied to build a classification model capable of handling non-linear data and finding optimal hyperplanes as the separation boundary between classes. This study used 50 samples obtained from POTADS in Lampung Province, Indonesia. The results showed that PCA successfully extracted three main components that explained 80.2% of the variance of the data. The SVM model achieved an overall accuracy of 80.2%, with the highest classification success rate at Level 1 (100%) and Level 3 (75%), while Level 2 had some classification errors due to a wider variation in sample characteristics. This study implies that the resulting model can be used to identify individuals at risk of motor difficulties, allowing for earlier and targeted behavior. In addition, the results of this study can be a reference for medical practitioners and educators in developing therapy and education strategies that are more in line with the needs of each individual.