CALL FOR PAPERS: SIGKDD Explorations Special Issue on Learning from Imbalanced Datasets
Guest editors: Nitesh Chawla (Customer Behavior Analytics, CIBC), Nathalie Japkowicz (University of Ottawa), Aleksander Kolcz (America Online, Inc.) SIGKDD Explorations is the official newsletter of ACM SIGKDD (Special Interest Group on Knowledge Discovery and Data Mining). This special issue invites submissions concerned with learning from imbalanced data. Many real-world problems are characterized by imbalanced learning data, where at least one class is under-represented relative to others, and/or where the data available for some of the classes does not reflect the true underlying distribution. In addition, there can be non-uniform costs associated with different types of errors or examples in the data. Important practical applications include: - fraud/intrusion detection - risk management - medical diagnosis/monitoring - bioinformatics - text categorization - personalization of information - direct marketing We invite contributions addressing the issues related to solving data mining involving imbalanced datasets. Relevant topics include (but are not limited to): - sampling (under-, over-, progressive, active) - post-processing of learned models - accounting for class imbalance via inductive bias - one-sided learning - handling uncertainty of target distribution and misclassification costs - handling varying amounts (class dependent) of label noise - handling the varying cost of procuring examples Submissions should be made to [EMAIL PROTECTED], preferably in a PDF format and should not exceed 8-10 pages. In addition, please email the abstract in text-format. Detailed formatting instructions are available from http://www.acm.org/sigkdd/explorations/instructions.htm. Submissions will be reviewed externally. Important Dates: Submissions January 30, 2004 Reviews due back to authors March 5, 2004 Camera-ready due April 5, 2004
