Research Article
Chinese Personal Name Disambiguation Based on Clustering
Table 5
Comparison of feature weight calculating approach (with highest
score).
| Feature weights | Document | Paragraph | Precision | Recall | score | Precision | Recall | score |
| Bool (N + NameEx) | 86.28% | 93.99% | 89.8% | 90.66% | 94.59% | 92.51% | Tf (N + NameEx) | 89.90% | 93.44% | 91.59% | 91.00% | 93.39% | 92.11% | Tf-idf (N + NameEx) | 91.07% | 94.22% | 92.58% | 91.29% | 93.12% | 92.15% | Entropy (N + Df1 + NameEx) | 90.54% | 92.49% | 91.47% | 89.63% | 86.88% | 88.04% |
|
|