Results - GeeUnit/hw5-team07 GitHub Wiki

CLEF 2012 Alzheimer's Test Set

The following table shows the tuning process on the K-Candidate values for the pipeline in the top row of the table on the Original (QA4MRE-2012_BIOMEDICAL_GS.xml) data set: images/KTuning-Original.png

CLEF 2013 Blind Set

The following table shows the tuning process on the K-Candidate values for the pipeline in the top row of the table on the Blind data set: images/KTuning-Blind.png

NOTES

We found that PMI scoring performed much better in the case of the datasets (with c@1 scores approaching north 0.5). However, it was less flexible, and had very poor runtime performance.