Machine Learning for Gravitas-Based Disease Prediction: CNN vs. Random Forest

The abstract highlights the main conclusions of the model's effectiveness for forecasting "Gravitas ratings." With remarkable precision values, this model exhibits a high level of classification efficacy and accuracy. Precision varies between 82.07% to 93.33% throughout this Gravitas rating class, demonstrating the model's capacity to produce accurate positive predictions. The model can properly detect positive examples, as shown by the recall values, which vary from 87.38% to 88.89%. The F1 scores, at an average of 87.97%, represent a harmonious combination of recall and precision. The support proportions also show good agreement with these criteria, demonstrating the model's dependability across various rating classes. Strong overall performance is indicated by its macro-average precision, recall, & F1-Score values (88.21%, 87.92%, & 88.01%, respectively). A collection with 4,208 photos for estimating disease severity. Our model's outstanding average accuracy of 87.92% highlights its efficacy in disease classification and prediction tasks. The average weighted metrics (88.13 percent precision, 87.93 percent recall, and 87.97 percent F1-Score) highlight the model's strong cross-class generalization. Its overall consistency in identifying all cases is highlighted by the micro-average measures (87.93% precision, recall, & F1-Score). As a result, the model is a useful tool for this categorization assignment because it performs consistently and dependably in predicting Gravitas scores.