ENHANCED CLASSIFICATION ACCURACY USING A PROPOSED DEEP LEARNING MODEL: A COMPARATIVE ANALYSIS WITH VGG, INCEPTION, LSTM, AND GRU ARCHITECTURES

Abstract

This study suggests a new deep learning classification model and compares its performance with widely used architectures—VGG, Inception, LSTM, and GRU—based on metrics such as F1 score, Precision, Accuracy, and Recall. The system utilizes hash values generated from data blocks and stores them in CSV files for input into the classification pipeline. The proposed model yields a final F1 score of 0.9819, accuracy of 0.982, and minimal classification latency of 0.0100 seconds, which is significantly better than the other models, according to testing results.