Integration Strategy Is a Key Step in Network-Based Analysis and Dramatically Affects Network Topological Properties and Inferring Outcomes

Table 6

The performance of the detection of disease genes using RWR.

Strategy	Rank	Nodes	Rank ratio
Naive Bayes	2065.86	14840	0.1392
Logistic Regression	2113.31	14890	0.1419
SVM	2133.86	14895	0.1433
Union	2146.21	14936	0.1437
Random Forest	2136.90	14860	0.1438
Bayesian Networks	2139.83	14869	0.1439
J48	2370.10	14570	0.1627
Random Tree	2697.93	14486	0.1862
3-Vote	1306.61	5558	0.2351
2-Vote	2245.89	9548	0.2352
Intersection	123.5	507	0.2436

Note: Rank indicates the average rank of the nonseed genes in several repeated experiments; the number of repetitions depended on the number of remaining genes. The rank ratio reveals the average rank divided by the total number of nodes in each network. The rank ratio was used to evaluate whether the performance of the integration strategy was outstanding. The smaller the scale is, the better the integration strategy is.