PubMed IDs and scores of the relevant training examples.
Started at
2009/01/18 23:43:10 GMT
Time at which query was started
Finished at
2009/01/18 23:44:26 GMT
Time at which this file was written.
Feature score method
scores_bgfreq
Name of the method used to calculate feature scores.
Docstring for the method: The prior is 'background frequency' successes, out of 1 total occurrence
in each class.
Min Document Frequency
2
We exclude features occurring fewer than this many times
in the data
Base score
-14.7415188786
The log likelihood ratio of an empty article (one in
which every feature failed to occur).
Prior score
-12.1288092152
The log of the prior probability ratio for
an article being relevant versus irrelevant (added to log likelihood ratio
to obtain the final score). Equals the logit of the estimated
prevalence of relevant articles in Medline (which may be estimated
from the input size or specified separately).
Limit
1000
The maximum number of results to include.
Threshold
-30.0
Default Naive Bayes classification threshold is
zero. This threshold is the minimum log probability ratio for
predicting an article to be relevant.
Feature Statistics
Quantity
Relevant Docs
Irrelevant Docs
Number of documents
91
17031885
Number of selected, occurring features
741
41209
Total occurrences of selected features
1943
231000504
Selected features per Medline record
21.352
13.563
Of the considered feature types, 41209 features are selected out
of 43325 occurring at least once in training data. The aggressivity of
selection is 1.051.
The complete database lists 43326 potential features.
Features with high TF.IDF
Features with TF.IDF above 0.2 or 0.3 could make good keywords. TF.IDF is term frequency times
inverse document frequency, where we treat the set of input citations as a
single document