Here are the Weka packages listed that I have developed over time:
- collective classification
semi-supervised learning and collective classification algorithms
This package is based on work from the original collective classification project, which was a hack for WEKA 3.5.x. The package manager approach represents a clean approach which does not rely on overwriting classes anymore.
The legacy code and datasets are available from the download section as well.
allows you to convert Weka command-line strings into various Java code snippets
- Confusion matrix visualizations
various visualizations that can be saved/printed, e.g., text, table, heatmap
- GraphViz Tree visualization
uses GraphViz to visualize trees generated by, e.g., J48
similar to the GridSearch optimization algorithm, but allows optimization of 1-to-n parameters
- Natural language processing
makes use of the Stanford Parser for part-of-speech (POS) tagging
allows recording of sound, webcam, screen feeds, to be combined in a screencast (and uploaded, e.g., to YouTube)
- Viewer for serialized models
adds a tab to the Weka Explorer that allows user to display content of Java serialized files
For a full list of Weka plugins, go to my curated list of projects on github.