Table 5 Performance comparison of the proposed BiLSTM model with other machine learning models.
From: An IoT-enabled AI framework for sustainable product design optimizing eco-efficiency using BiLSTM
Model | Accuracy (%) | F1-Score (%) | Sensitivity (%) | Specificity (%) |
|---|---|---|---|---|
Decision Tree | 85.4 ± 0.44 | 83.9 ± 0.37 | 83.5 ± 0.52 | 86.2 ± 0.48 |
Random Forest | 88.7 ± 0.41 | 87.0 ± 0.35 | 86.8 ± 0.49 | 89.1 ± 0.45 |
SVM | 86.2 ± 0.45 | 85.3 ± 0.37 | 85.0 ± 0.51 | 87.4 ± 0.47 |
Naive Bayes | 82.9 ± 0.49 | 81.5 ± 0.40 | 81.2 ± 0.55 | 84.3 ± 0.51 |
KNN | 84.1 ± 0.51 | 82.8 ± 0.42 | 82.5 ± 0.53 | 85.0 ± 0.50 |
XGBoost | 90.5 ± 0.33 | 89.6 ± 0.27 | 89.3 ± 0.44 | 91.2 ± 0.41 |
CNN | 92.3 ± 0.28 | 91.4 ± 0.23 | 91.1 ± 0.40 | 93.0 ± 0.38 |
BiLSTM (Proposed) | 97.6 ± 0.21 | 96.1 ± 0.22 | 95.9 ± 0.28 | 98.2 ± 0.25 |
