Comparative Performance Evaluation of SVM, KNN, and Decision Tree Algorithms in Facial Recognition Tasks Using PCA

Face recognition is a widely used technology in various fields, including security, user authentication, and video analytics. This study evaluates the comparative performance of classification algorithms—Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Decision Tree—in detecting faces. To enhance processing efficiency, Principal Component Analysis (PCA) is employed as a feature extraction and dimensionality reduction method, as several studies have shown that PCA can help reduce the dimensionality of large datasets, thereby improving both the efficiency and accuracy of classification models. The dataset used consists of 77,658 images categorized into two classes: owner and strangers. The models are trained using hyperparameter tuning and evaluated through k-fold cross-validation repeated over 10 iterations to ensure a robust assessment by averaging the results across key performance metrics. Performance is measured using main metrics such as Accuracy, Precision, Sensitivity, and F1-Score. Experimental results show that the Decision Tree algorithm outperforms SVM and KNN, achieving the highest accuracy of 0.978189 on fold 3. For the owner class, it achieved a precision of 0.912543, a sensitivity of 0.680519, and an F1-Score of 0.779416. Meanwhile, for the strangers class, it achieved a precision of 0.981101, a sensitivity of 0.996067, and an F1-Score of 0.988527. This study provides valuable insights into the effectiveness of different classification methods in facial detection systems and highlights the Decision Tree algorithm as the most effective in handling facial recognition tasks efficiently, despite its challenges in predicting owner face data. It is hoped that this research can be further developed by using larger and more diverse datasets, including variations in facial expressions, viewpoints, and lighting conditions, as well as improvements in owner face identification and the implementation of real-time face recognition systems.