Research Article
Distance Variance Score: An Efficient Feature Selection Method in Text Classification
Table 3
The sparsity of sub-DTMs constructed by features selected by LS and DVS (unit: %).
| DBWorld data set | CNAE data set | | DVS | LS | | DVS | LS |
| 3721 | 95.01 | 856 | 99.20 | 2721 | 93.76 | 97.28 | 656 | 99.00 | 99.01 | 1721 | 91.04 | 98.00 | 456 | 98.60 | 98.81 | 721 | 83.16 | 98.10 | 256 | 97.73 | 98.68 | 521 | 79.47 | 98.03 | 156 | 96.60 | 99.78 | 321 | 73.01 | 98.40 | 56 | 92.80 | 99.90 | 121 | 61.23 | 98.44 | | | | 21 | 49.70 | 98.44 | | | |
|
|