bst-mug / n2c2 / 252

Builds Branch Commit Type Ran Committer Via Coverage
252 master Remove unused weka `MultiTokenizer` This was part of an unsuccessful SVM experiment and is not longer used. push michelole travis-ci pending completion   set done
251 master Merge pull request #77 from michelole/issue-73 Issue 73 push web-flow travis-ci pending completion   set done
249 master Properly truncate short narratives during prediction When training, we learn `maxLength`, which is the largest number of sentences found on a single narrative. When predicting, we initialize a feature vector with the same size as the document to... push michelole travis-ci pending completion   set done
248 develop loadFeaturesForNarrativeDebug little change... push mtothak travis-ci pending completion   set done
247 master Log average accuracy push michelole travis-ci pending completion   set done
246 master Reset `trainCounter` when training new examples `initializeNetworkFromFile` will always try to get a file with `trainCounter` set to zero, which fails when the class is reused for multiple datasets. Therefore, reset the `trainCounter` to zero wh... push michelole travis-ci pending completion   set done
245 master Make prediction 13 times faster `net.output` is the most expensive call during prediction, but it's only dependent on the features extracted form a given patient text. Therefore, there's no need to call it once per criterion; instead, predict th... push michelole travis-ci pending completion   set done
244 master Introduce `FalseAnalysisRunner` (fixes #68) push michelole travis-ci pending completion   set done
243 master Use only the split when testing on trained data (fixes #24) push michelole travis-ci pending completion   set done
242 master Add main method to generate training curves (fixes #24) push michelole travis-ci pending completion   set done
241 master Remove unused dependencies (refs #72) push michelole travis-ci pending completion   set done
240 master Upgrade lucene to 7.5.0 (refs #72) Also, remove `lucene-core` explicit dependency, as it's internally required by `lucene-analyzers-common`. push michelole travis-ci pending completion   set done
239 master Add an active unit test for neural networks This allows us to detect unused code more easily, detect regressions when updating libraries, as well as testing on different JDKs. push michelole travis-ci pending completion   set done
238 master Merge pull request #71 from michelole/issue-67 Issue 67 push web-flow travis-ci pending completion   set done
237 issue-67 Standardize `NNClassifierRunner` (fixes #67) Now that we are able to use NN classifiers within our framework, standardize our runner method. Pull #71 web-flow travis-ci pending completion   set done