Figure 1.
Related Work Study Plan
Figure 2.
Dataset
Figure 3.
Dataset Preprocess
Figure 4.
Category count
Figure 5.
Dataset Translation rate
Figure 6.
Logistic Regression
Figure 7.
Support Vector Machine
Figure 8.
K-Nearest Neighbor
Figure 9.
GCN Work Flow
Figure 10.
mBERT
Table 1.
Imbalanced Dataset without Translation
| Type | Algorithm | F1 Score (%) | Time (s) | Kappa (%) |
|---|---|---|---|---|
| BOW | Bernoulli Naive Bayes | 88 | 0.08 | 27 |
| Support Vector Machine | 88 | 1281.27 | 45 | |
| Logistic Regression | 89 | 2.35 | 45 | |
| K-Nearest Neighbor | 86 | 0.03 | 0 | |
| TF-IDF | Bernoulli Naive Bayes | 88 | 0.07 | 27 |
| Support Vector Machine | 89 | 469.69 | 47 | |
| Logistic Regression | 88 | 1.4 | 42 | |
| K-Nearest Neighbor | 86 | 0.03 | 1 | |
| Word2Vec | Bernoulli Naive Bayes | 74 | 0.13 | 26 |
| Support Vector Machine | 86 | 747.66 | 0 | |
| Logistic Regression | 86 | 2.57 | 14 | |
| K-Nearest Neighbor | 86 | 0.05 | 4 |
Table 2.
Imbalanced Dataset with Translation
| Type | Algorithm | F1 Score (%) | Time (s) | Kappa (%) |
|---|---|---|---|---|
| BOW | Bernoulli Naive Bayes | 86 | 0.35 | 25 |
| Support Vector Machine | 86 | 4290.71 | 25 | |
| Logistic Regression | 87 | 5.2 | 29 | |
| K-Nearest Neighbor | 86 | 0.06 | 0 | |
| TF-IDF | Bernoulli Naive Bayes | 86 | 0.17 | 25 |
| Support Vector Machine | 87 | 1698.51 | 22 | |
| Logistic Regression | 87 | 2.46 | 28 | |
| K-Nearest Neighbor | 86 | 0.06 | 0 | |
| Word2Vec | Bernoulli Naive Bayes | 73 | 0.12 | 26 |
| Support Vector Machine | 86 | 444.38 | 0 | |
| Logistic Regression | 86 | 4.3 | 6 | |
| K-Nearest Neighbor | 86 | 0.05 | 10 |
Table 3.
Balanced Dataset without Translation
| Type | Algorithm | F1 Score (%) | Time (s) | Kappa (%) |
|---|---|---|---|---|
| BOW | Bernoulli Naive Bayes | 80 | 82 | 0.35 |
| Support Vector Machine | 82 | 83 | 4290.71 | |
| Logistic Regression | 82 | 83 | 5.2 | |
| K-Nearest Neighbor | 16 | 54 | 0.06 | |
| TF-IDF | Bernoulli Naive Bayes | 86 | 85 | 0.17 |
| Support Vector Machine | 85 | 85 | 1698.51 | |
| Logistic Regression | 82 | 83 | 2.46 | |
| K-Nearest Neighbor | 7 | 52 | 0.06 | |
| Word2Vec | Bernoulli Naive Bayes | 62 | 66 | 0.27 |
| Support Vector Machine | 68 | 72 | 1645.81 | |
| Logistic Regression | 70 | 73 | 7.39 | |
| K-Nearest Neighbor | 68 | 72 | 0.07 |
Table 4.
Balanced Dataset with Translation
| Type | Algorithm | F1 Score (%) | Time (s) | Kappa (%) |
|---|---|---|---|---|
| BOW | Bernoulli Naive Bayes | 71 | 73 | 0.17 |
| Support Vector Machine | 75 | 78 | 2845.36 | |
| Logistic Regression | 75 | 76 | 3.2 | |
| K-Nearest Neighbor | 18 | 54 | 0.06 | |
| TF-IDF | Bernoulli Naive Bayes | 76 | 74 | 0.32 |
| Support Vector Machine | 75 | 76 | 1207.85 | |
| Logistic Regression | 74 | 75 | 1.53 | |
| K-Nearest Neighbor | 26 | 52 | 0.07 | |
| Word2Vec | Bernoulli Naive Bayes | 65 | 69 | 0.27 |
| Support Vector Machine | 67 | 72 | 1912.06 | |
| Logistic Regression | 69 | 72 | 5.97 | |
| K-Nearest Neighbor | 65 | 71 | 0.07 |