<![CDATA[This research implements the Decision Tree algorithm in a Smart Glove to classify hand signals representing numbers 1 to 5 in the Indonesian Sign Language System (SIBI). The system utilizes flex and gyro sensors to capture hand movements, which are then processed and classified using the Decision Tree algorithm. Training data was collected from multiple trials, resulting in an accuracy of 79% across 50 trials. The model's performance evaluation yielded precision, recall, and F1-score values ranging between 80% and 90% for each number class. The best performance was achieved with number 1, reaching 90% in precision, recall, and F1-score. However, areas for improvement were identified in precision and recall for numbers 2 and 4. Although the results are adequate, this study highlights the need for further development in enhancing model accuracy, particularly by increasing training data and refining the algorithm. The Smart Glove is expected to aid communication for individuals with hearing disabilities and holds potential for future expansion to recognize more complex gestures.]]>
Copyrights © 2024
