An Empirical Study of Learning from Imbalanced Data Using Random Forest

Authors:
Taghi M. Khoshgoftaar;Moiz Golawala;Jason Van Hulse
Affiliations:
-;-;-
Venue:
ICTAI '07 Proceedings of the 19th IEEE International Conference on Tools with Artificial Intelligence - Volume 02
Year:
2007

Citing 0
Cited 11

Aggregating performance metrics for classifier evaluation

IRI'09 Proceedings of the 10th IEEE international conference on Information Reuse & Integration
An empirical investigation of filter attribute selection techniques for software quality classification

IRI'09 Proceedings of the 10th IEEE international conference on Information Reuse & Integration
Data mining for credit card fraud: A comparative study

Decision Support Systems
Predicting high-risk program modules by selecting the right software measurements

Software Quality Control
Software measurement data reduction using ensemble techniques

Neurocomputing
Leveraging editor collaboration patterns in wikipedia

Proceedings of the 23rd ACM conference on Hypertext and social media
Color fused multiple features for traffic sign recognition

Proceedings of the 4th International Conference on Internet Multimedia Computing and Service
An Empirical Investigation of Filter Attribute Selection Techniques for High-Speed Network Traffic Flow Classification

Wireless Personal Communications: An International Journal
Discovering health-related knowledge in social media using ensembles of heterogeneous features

Proceedings of the 22nd ACM international conference on Conference on information & knowledge management
Identification of micro RNA biomarkers for cancer by combining multiple feature selection techniques

Journal of Computational Methods in Sciences and Engineering
Training and assessing classification rules with imbalanced data

Data Mining and Knowledge Discovery

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper discusses a comprehensive suite of experi- ments that analyze the performance of the random forest (RF) learner implemented in Weka. RF is a relatively new learner, and to the best of our knowledge, only preliminary experimentation on the construction of random forest clas- sifiers in the context of imbalanced data has been reported in previous work. Therefore, the contribution of this study is to provide an extensive empirical evaluation of RF learn- ers built from imbalanced data. What should be the rec- ommended default number of trees in the ensemble? What should the recommended value be for the number of at- tributes? How does the RF learner perform on imbalanced data when compared with other commonly-used learners? We address these and other related issues in this work.