Attribute Importance Ranking List


 

For each of the 15 attributes we have used, we have counted the percentage of the top four nodes (in all the random forest trees) that use this attribute. These are summarized in the table below. (The first column contains the position index of the attribute within our feature set.)

 

Table S3. Attribute Importance

Attribute No.

Dataset

Importance  ranking

 (Counting Way, n=4)

3

Gene Expression

0.1868

15

Protein Expression

0.1302

2

Domain-Domain Interaction

0.1154

5

GO Biological Process

0.1053

8

TAP Mass

0.0800

13

Synthetic Lethal

0.0739

6

GO Cellular Component

0.0570

4

GO Molecular Function

0.0519

7

Y2H

0.0427

9

HMS_PCI Mass

0.0425

14

Syn-expression

0.0385

12

Gene Co-occur

0.0317

10

Gene Neighborhood

0.0211

1

Protein-DNA Binding

0.0178

11

Gene Fusion

0.0052