One important way that Science progresses and knowledge accumulates is by the discovery and explanation of exceptions. Several historical examples, widely quoted in the popular press, illustrate that it is often the chance discovery and/or explanation of outliers which leads to a new understanding of the world around us: In 1901, Fred McKay, then a fresh dentistry graduate noticed that children in Colorado Springs, USA, had unusually healthy teeth. Eventually he isolated the cause to the presence of fluoride in the local water supply. Fluoridation of water is now a common practice in all modern cities and nearly every toothpaste sold in the market contains fluoride as an active ingredient. A similar example, of much contemporary interest, is the chance discovery of an explanation of the climate phenomenon known as El Nino. A British meteorologist, Sir Gilbert Walker, noticed that the outlier behaviour of sea surface pressure difference between Darwin, Australia and Tahiti could help predict rainfall patterns in different parts of the world and was related to El Nino. While these two examples suggest that outliers act as triggers which lead to new discoveries, the explanation of known outliers can also be used to validate theories. In fact one of the earliest confirmation of Einstein’s General Theory of Relativity was its ability to explain the anomalous orbit of planet Mercury which hitherto could not be exactly explained on the basis of Newtonian mechanics. On the financial side, there is also much interest in "extreme events" since such events have major impacts on financial markets as witnessed during July and August of 2007 and as recently as January of 2008. Outliers continue to play a major role in our understanding of natural and social systems. One common reason for detecting outliers is to mediate their effect on statistical estimators through a process known as robustification. There are several ways of robustifying an estimator and they all involve first identifying the “outliers” and then carrying out a transformation on them. For example, ordered data can be “winsorized at 90%” by equating the bottom 5% of the values to the fifth percentile and the upper 5% of the data to the ninety-fifth percentile. . The use of outlier detection techniques as a tool to “discover” new knowledge and gain fresh insights about the underlying data generating process has attracted increasing attention as we spiral into a “data rich environment.” Increasing volumes of data suggest that we design algorithms and build information systems where outliers can be detected and dealt with minimal human intervention. Scope The purpose of the special issue is to bring together current insights, ideas and application of outliers as a vehicle of “knowledge discovery.” We invite theoretical and application papers in the following suggested (but not limited to) topics: Definition of outliers in context. Efficient algorithms for outlier detection. Validation of outliers in high dimensional space. The use of information theory to understand outliers. Applications and case studies of outlier detection for knowledge discovery. Outlier detection as classification of imbalanced data sets. The discovery and validation of low support and high confidence association rules. Outlier detection for spatial, temporal and sequential data. Submission Guidelines Submission information can be found on the journal web page by clicking on "Instructions for Authors" on the sidebar. There are no specific page length restrictions. Authors should submit their manuscripts using DMKD's online system. To submit your article, go to the online manuscript submission page, select "submit a manuscript", create a user account, and, when prompted, choose "Special Issue: Outlier Detection for Knowledge Discovery" as the article type. Please feel free to contact the guest editors with any questions. Important Dates Submission Deadline:    July 18, 2008 Expected Author Notification:    October, 2008 Expected submission of final manuscripts:    February, 2009 Special Issue Guest Editors Sanjay Chawla, University of Sydney, chawla@it.usyd.edu.au David Hand, Imperial College, d.j.hand@imperial.ac.uk Vasant Dhar, Stern Business School, New York University, vdhar@stern.nyu.edu