<![CDATA[Pneumonia is one of the leading causes of death in children, so early detection is very important to increase the chances of recovery. This study aims to develop a hybrid model based on artificial intelligence to detect pneumonia in children by combining X-ray images and clinical data. The system is built in two main stages: image feature extraction using Convolutional Neural Network (CNN) based on MobileNetV2 architecture, and combined classification of visual and clinical features using Support Vector Machine (SVM). The CNN model successfully classified children's lung conditions with an accuracy of 89.92% and a recall of 0.95 in the pneumonia class, indicating excellent detection sensitivity. The SVM model built using clinical data also showed competitive performance with an accuracy of 90%, confirming the important contribution of clinical data in supporting diagnosis. The hybrid model combining visual feature vectors (128 dimensions) and 12 clinical features showed very high performance with an accuracy of 98% on the test data, and precision and recall above 0.98 in both classes. Principal Component Analysis (PCA) visualization shows a clearly separated class distribution, supporting the reliability of the multimodal representation used. The results of this study indicate that the integration of visual and clinical data significantly improves the accuracy and sensitivity of the model, making it a potential candidate as an AI-based pneumonia diagnostic tool. This model is very relevant to be applied in areas with limited medical personnel, to support faster and more accurate early detection. Further research is recommended to test the model's performance on more heterogeneous external data to ensure generalization capabilities.]]>
