Classification in E-Procurement - University of Reading
Classification in E-Procurement - University of Reading
Classification in E-Procurement - University of Reading
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Methods Tried<br />
K-nearest Neighbour (prelm<strong>in</strong>ary tests show K best at 5)<br />
Rocchio – equal balalnce <strong>of</strong> negative and positive prototypes<br />
Naïve Bayes – Bernoulli model<br />
Support Vector Mach<strong>in</strong>e – l<strong>in</strong>ear models<br />
Two Null hypotheses – as control (random or most <strong>of</strong>ten used)<br />
Tested on Reuters data set and on PO data<br />
Performance assessed by F measure – mean <strong>of</strong> precision / recall<br />
Macro averaged (across all classes)<br />
Micro averaged (sum <strong>of</strong> each class)<br />
p c =<br />
TPc<br />
TP c + FPc<br />
r c =<br />
TPc<br />
TP c + FNc<br />
<strong>Classification</strong> <strong>in</strong> E-<strong>Procurement</strong>- 9<br />
Presented at CIS 2012 © Dr Richard Mitchell 2012