Skip to main content
. 2016 Oct 6;17(Suppl 13):335. doi: 10.1186/s12859-016-1220-5

Fig. 7.

Fig. 7

Performance evaluation results for DEGTwitter as new features are added to the model. Left: Mean 10-fold cross validation results with errors bars indicating +/- 1 standard deviation. Actual values in table below Right: Performance results on the test set. Performance on the test set steadily increases with the expanding feature set. Biggest gains are achieved when adding topic context and enriching the DEGTwitter with features from DEGSIder