|
| Reference | CI techniques | Performance | Datasets |
|
| [118] | Markov random field | Specificity: 45%, sensitivity: 64% | Functional probability of each protein |
| [119] | Network flow based algorithm | Accuracy: 10–90% | Structure of protein interactive maps |
| [120] | Neighbor based techniques | Precision: 0.9-1.0, recall: 98% | Label 1 and label 2 neighbors |
| [121] | Association analysis based method | Accuracy: 93% | H confidence, adjacency matrix |
| [122] | Naïve Bayes classifier | Precision: 49%, recall: 62%, MCC: 0.37 | PPI data |
| [123] | RWR with -NN | Accuracy: 58–73% | Neighborhood features |
| [124] | Time sequenced subnetwork | Significant module: 95.95% | Integrating the gene expression data and PPI data |
| [125] | Gibbs sampling based bootstrapping | TP/FP: 0.5 to 1.5 | Interaction and annotation data |
| [126] | Network based approach | Precision: 54.83%,
-score: 43.74% | Function-function correlation |
| [127] | Neighborhood majority voting system | Precision: 67.3%,
recall: 40.30% | Diffusion state distance (DSD) |
|
