Research Article
A Dynamic Model for Imputing Missing Medical Data: A Multiobjective Particle Swarm Optimization Algorithm
Table 5
The comparison of the proposed model of imputation with EM algorithm for ATLL patients’ data.
| | Algorithm name | Sensitivity (%) | Specificity (%) | Accuracy (%) | PPV+ (%) | PPV− (%) | F-measure (%) |
| | Delete missing | 47.00 | 40.60 | 45.95 | 47.1 | 38.95 | 45.47 | | Mean algorithm | 47.37 | 51.20 | 53.77 | 44.77 | 49.83 | 43.49 | | Expectation maximization | 62.57 | 69.25 | 70.23 | 64.44 | 65.35 | 61.31 | | MICE algorithm | 46.16 | 49.28 | 53.37 | 45.92 | 46.88 | 43.09 | | missForest algorithm | 58.30 | 62.65 | 64.65 | 58.15 | 61.00 | 56.09 | | Proposed algorithm | 86.15 | 82.4 | 86.75 | 83.57 | 84.67 | 83.50 |
|
|
