Synergistic CNN-Random Forest Model for Signature Recognition and Classification

The abstract performance characteristics of a signature classifying hypothesis, including precision, recall, F1-Score, encouragement, support percentage, and accuracy for two different signature classification levels. To take a Signature Classification stage one, the proposed approach achieves remarkable accuracy, recall, and F1-Score ratings of 99.87%, 99.50%, & 99.68%, accordingly. These metrics show how amazingly well the algorithm can classify situations in this category. The robustness of the model is demonstrated by the support column, which has 4020 instances classified as Signature Classification Stage 1. With a significant value of 0.47, the support percentage indicates the class's contribution to the entire dataset. In a similar vein, Signature Classification Stage 2 exhibits excellent performance characteristics, with 99.77 percent accuracy, recall, and F1-Score. There is a greater support rate in this class since 4510 instances give 0.52 to the total dataset. The model demonstrates its ability to identify signals with a stunning 99.64% Small Average accuracy, which is the average precision over all cases. Both signature levels performed in balance, as indicated by the weighted average & macro average metrics, which are predicted to be 99.73% for F1-Score, 99.82% for accuracy, & 99.64% for recall. This demonstrates how the model succeeds in maintaining a high level of accuracy throughout the entire dataset in addition to predicting specific classes. With a 99.6483% overall accuracy, it is shown that the notion of signature classification helps classify signatures at different levels.