Research Article

Chinese Personal Name Disambiguation Based on Clustering

Table 4

Comparison of feature selection (tf-idf).

FeaturesDocumentParagraph
PrecisionRecall scorePrecisionRecall score

NE80.15%92.24%83.32%77.74%88.84%80.36%
N81.38%93.74%84.59%82.49%92.85%85.15%
N + V80.10%93.83%83.83%80.93%93.11%83.86%
N + Df182.70%92.76%85.45%82.77%90.40%84.75%
NE + NameEx89.32%92.55%90.87%86.81%88.94%87.79%
N + NameEx91.07%94.22%92.58%91.29%93.12%92.15%
N + V + NameEx89.57%94.31%91.82%90.55%93.45%91.92%
N + Df1 + NameEx91.09%93.25%92.11%90.28%90.65%90.40%