| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.txt | 2015-02-13 | 1.3 kB | |
| meka-release-1.7.5.zip | 2015-02-13 | 15.9 MB | |
| meka-javadoc-1.7.5.zip | 2015-02-13 | 716.4 kB | |
| meka-src-1.7.5.jar | 2015-02-13 | 492.3 kB | |
| Tutorial.pdf | 2015-02-13 | 446.1 kB | |
| meka-1.7.5-tests.jar | 2015-02-13 | 39.8 kB | |
| meka-1.7.5.jar | 2015-02-13 | 364.3 kB | |
| Totals: 7 Items | 17.9 MB | 0 | |
See the Tutorial.pdf for detailed information on obtaining, using and extending MEKA. For a list of included methods and 'quick-start' command line examples, see: http://meka.sourceforge.net/methods.html Release Notes ------------- This is a relatively minor release. There are no fundamental changes to the internals. However, many classifiers have been revised in some way since the last release, and there are several new classifiers: - CT: classifier trellis - CDT: conditional depndency trellis - HASEL: a RAkEL-like algorithm, that partitions labelsets based on a dataset defined hierarchy Also, there have been several small but helpful changes to the GUI: - An indicator of when an experiment is in progress - Label variables are set in *bold* to distinguish from the input/attribute variables - Cleaner output - Bookmarks in the load-file dialog. - Minor bug fixes - Default classifiers have been set sensibly (not ZeroR for everything as before) - Some 'missing' classifier options are now present for manipulation And a lot of minor changes throughout - Sensible default classifiers for the command line - Updated tutorial - Improved documentation all-around (tutorial, javadoc and code comments and list of methods with examples: http://meka.sourceforge.net/methods.html)