Research Article
Identifying Potential Clinical Syndromes of Hepatocellular Carcinoma
Using PSO-Based Hierarchical Feature Selection Algorithm
Table 3
The predicting performance of the optimal feature subsets obtained from different feature selection methods.
| | Approaches | Dimension of the optimal feature subset | MSE | RMSE | MRPE (%) | Time (second) |
| | PSOHFS | 24 (syndromes) | 0.1622 | 0.4027 | 1.0700 | 3.0108 | | CFM (top 15%) | 22 (symptoms) | 14.4575 | 3.8023 | 11.8907 | 2.8510 | | CFM (top 30%) | 45 (symptoms) | 6.2611 | 2.5022 | 7.8632 | 4.8010 | | PWM (100 iterations) | 92 (symptoms) | 3.2268 | 1.7963 | 5.5645 | 8.8760 | | PWM (200 iterations) | 89 (symptoms) | 2.7516 | 1.6588 | 5.2351 | 8.7390 |
|
|
