RSS 2.0
Critical Assessment of Information Extraction in Biology - data sets are available from Resources/Corpora and require registration.

BioCreative II.5

Raw participant data [2010-01-14]

This package contains the raw data for the BioCreative II.5 online challenge submitted by the participants. Inside the archive are two folders, one called "online", the other "offline", i.e., grouped by the submission type (sent offline or via an Annotation Server). This data is the basis for generating the official results, calculated with the BioCreative evaluation library.

The offline file naming convention is as follows:
[teamID]_[taskID]_[runID]_(train|test).tsv
The online file naming convention is similar:
t[teamID]_s[serverID]_(train|test)_[taskID].tsv

The content is tab-separated text using the formatting guidelines of the challenge; results may contain ranks or not, but will always contain confidence scores.

Additionally, as reference, an Excel file to make the association between team IDs and the official participants is provided, which can be found online here, too.

Downloads