0
I Use This!
Very Low Activity

Commits : Listings

Analyzed 1 day ago. based on code collected 1 day ago.
May 30, 2023 — May 30, 2024
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Remove the commons-logging library More... over 6 years ago
Remove the netlib library More... over 6 years ago
Remove the trove library More... over 6 years ago
Remove the Sparse Matrix library. If we need high performance sparse matrix operations later we should have a look at https://github.com/fommil/matrix-toolkits-java More... over 6 years ago
Remove the jwnl library. More... over 6 years ago
Make sure the test classes are removed as well when cleaning. More... over 6 years ago
Remove the commons-codec library More... over 6 years ago
Remove the BeanShell library More... over 6 years ago
Remove the arpack library. Not even sure what this was, maybe related to https://github.com/fommil/netlib-java More... over 6 years ago
It seems the ngrams work as expected now as well. More... over 6 years ago
Also change the sparse extractor and its test to not set the type name used for the external feature name to blank any more if it was not blank in the specification and matches the effective type from the PR. More... over 6 years ago
Change the way how annotation types are handled in feature names. This has now been changed for the dense extractions and also needs to get changed for the sparse extraction: the annotation type part in the feature name will always be just what is declared in the attribute specification. If a type is set in the specification, it will NOT get set to blank any more if the instance type specified in the PR matches it! More... over 6 years ago
Looks like we are getting the correct dense instance representations now. More... over 6 years ago
OK, start working on adding the corpus representation. Needs some refactoring since this is a totally new way to do things. More... over 6 years ago
Make unit tests work for the basic extraction bits. More... over 6 years ago
Add file for testing dense, so far only a copy. Also add Netbeans project since automatic projects cannot deal with separate main and test directories with ant. Probably can use automatic again once this is fully mavenized. More... over 6 years ago
Initial draft for a pom. But for now we cannot compile using Maven since we still depend on libraries which are not in Maven. More... over 6 years ago
Change layout of source directory to follow Maven conventions. More... over 6 years ago
Add dense feature extraction and instance representations. More... over 6 years ago
Just a tiny change to start the branch. More... over 6 years ago
Fix issue #32, throw exception when instance without target value is found. More... over 6 years ago
Comment typos. More... over 6 years ago
Exception chaining More... over 6 years ago
Correct trainer class for MALLET_CL_NAIVE_BAYES More... over 6 years ago
Better handling of escaping/cleaning strings when exporting to tsv/csv format. More... over 6 years ago
Better handling of missing values for number-coded nominal attributes when exporting to ARFF format. More... over 6 years ago
Make sure an empty string is properly escaped for ARFF data format and that null is equal to the empty string. More... over 6 years ago
Second step of refactoring: use new methods to implement the JSON exporter for classification/regression. More... over 6 years ago
First step of major refactoring of the JSON exporter. This refactors the sequence exporter and splits the task up into converting a feature vector to a string, a target to a string and a sequence example to a string. More... over 6 years ago
Support option -S to output nominal values coded as number and also targets as the original string if the output format is TSV. More... over 6 years ago