Figure 4
The chart is a grouped vertical bar graph with eight model categories along the horizontal axis (Models) labeled from left to right as “C N N (Res Net 18)”, “L S T M”, “CLIP”, “C N N plus L S T M”, “CLIP plus C N N”, “CLIP plus L S T M”, and “H I P M”. The vertical axis is labeled “Metrics” and ranges from 0 percent to 100 percent with ticks at every 10 percent. Each model group contains four colored bars representing “Precision” (pink), “Recall” (orange), “F 1 score” (blue), and “A U C” (green), with a matching color legend centered below the plot. The exact percentage for each metric is printed in black text at the top of its bar.zz For C N N (Res Net 18): the precision is 80.33 percent, the recall is 77.15 percent, the F 1 score is 75.89 percent, and the A U C is 92.61 percent. For L S T M: the precision is 47.79 percent, the recall is 59.74 percent, the F 1 score is 51.92 percent, and the A U C is 77.15 percent. For C L I P: the precision is 45.72 percent, the recall is 53.00 percent, the F 1 score is 48.00 percent, and the A U C is 62.51 percent. For C N N plus L S T M: the precision is 25.95 percent, the recall is 50.94 percent, the F 1 score is 34.38 percent, and the A U C is 53.44 percent. For C L I P plus C N N: the precision is 88.60 percent, the recall is 88.85 percent, the F 1 score is 88.40 percent, and the A U C is 93.56 percent. For C L I P plus L S T M: the precision is 88.22 percent, the recall is 82.77 percent, the F 1 score is 84.15 percent, and the A U C is 91.11 percent. For H I P M: the precision is 91.52 percent, the recall is 91.20 percent, the F 1 score is 91.20 percent, and the A U C is 95.36 percent.Performance comparison of different model combinations – CNN, LSTM, CLIP and their hybrid architectures (CLIP + CNN, CLIP + LSTM, CNN + LSTM and HIMP) – evaluated using precision, recall, F1 score and AUC. The bar chart visually highlights the impact of different model architectures on performance, with hybrid models demonstrating varying levels of effectiveness. The CLIP + LSTM model shows relatively lower performance, whereas CNN + LSTM and CLIP + CNN exhibit stronger results. The proposed HIMP model achieves the highest scores across all metrics, notably with an F1 score of 91.20% and an AUC of 95.36%, confirming its superiority over other combinations. Source: Authors’ own work