Red(uced)-RF, a new type of Random Forests that adopts dynamic data reduction and weighted upvoting techniques. Red-RF is favorably applicable to big data: it demonstrates an accurate and efficient performance while achieving a considerable data reduction w.r.t. dataset size.
Manuscripts available on IEEE Xplore:
H. Mohsen, H. Kurban, K. Zimmer, M. Jenne and M. Dalkilic. Red-RF: Reduced Random Forests using priority voting & dynamic data reduction. In IEEE BigData Congress'2015.
H. Mohsen, H. Kurban, M. Jenne and M. Dalkilic (2014). A New Set of Random Forests with Varying Dynamic Data Reduction and Voting Techniques. In IEEE DSAA'2014.
Code, README file, and a sample input file are available in Files/ directory above.
For inquiries, please contact us at hmohsen@imail,iu.edu (or @indiana.edu).
Features
- Data Reduction
- Classification
- Random Forests
- Weighted Voting
- Machine Learning
- Data Mining
- Big Data