Association (frequent pattern) mining is a fruitful and important area of study that seeks (a) to discover what items (objects) appear frequently together and (b) to predict, given a set of objects, what other objects are likely to occur. Association mining has been successfully utilized for web log analysis, analysis of customer buying, biomedical knowledge mining, among others. However, a key drawback in association mining is the users need to take additional steps to make the information actionable. This includes, in many cases, in detecting what differences resides between groups, such that any changes are actionable. Research in this area finds a home in a number of related aread, including discriminative pattern mining, contrast mining, emerging pattern mining, action rule mining and subgroup discovery. The motivation in organizing this workshop is to attract novel papers in the development and application of algorithms within the above fields and to encourage conversations and idea exchange between the subfields.
For the ease of use, when using Discriminative Mining, we use the term to include (but not limited to), discriminative pattern mining, contrast mining, emerging pattern mining, action rule mining and subgroup discovery.
While not comprehensive, topics of interest include:
• Discriminative mining and learning in big data
• Theory, applications, and core methods
• Discriminative mining in large datasets
• Predictive modeling/learning, clustering and data analysis that incorporates Discriminative mining
• Incremental and/or streaming discriminative mining
• Privacy/security in discriminative mining
• Visualization techniques for utility discriminative mining
• Applications of discriminative mining in healthcare, manufacturing, predictive and prescriptive maintenance, social media, etc.
The paper should be up to 10 pages using 2-column IEEE Computer Science Workshop paper format.