Skip to main content
. 2014 Nov 23;2014:bau113. doi: 10.1093/database/bau113

Table 3.

Method description of submitted runs

Subtask Run ID Method description
1 Run 1 RDE, 110 reference features, Logistic Regression, classification threshold = 0.16
1 Run 2 RDE, 110 reference features, Logistic Regression, classification threshold = 0.18
1 Run 3 RDE, 110 reference features, Logistic Regression, classification threshold = 0.14
2 Run 1 GO Rank, Hierarchy filtering, GO terms with the count over 2000 in the GOA database for ranking. classification threshold (Subtask1) = 0, filtering threshold = 6
2 Run 2 GO Rank, Hierarchy filtering, GO terms with the count over 500 in the GOA database for ranking, classification threshold (Subtask1) = 0, filtering threshold = 8
2 Run 3 GO Rank, Hierarchy filtering, GO terms with the count over 2000 in the GOA database for ranking, classification threshold (Subtask1) = 0.16, filtering threshold = 2

In the table, ‘classification threshold’ is the threshold of the Logistic regression classifier with 110 RDE features. The ‘filtering threshold’ is the number of n most relevant high-level GO classes to the sentence determined by the classifiers. If the highest ranked GO term by GORank is in the n classes, it will be selected as a positive result.