This year RuleML-2014 will be held in Prague on Aug 18-20. For the first time it will include a special track called “Learning Business Rules from Data”. As a member of the organizing committee, I posted the proper announcement here. It promises to become a very interesting event when the decision management practitioners meet their academic partners. The topics of the track will address (among others) extraction of business rules from sets of fuzzy, uncertain and possibly conflicting rules learned from data and bridging the gap between rules as “correlations” in the data and rules that can be used in business rule management systems.
Our Rule Learner is a good practical tool that integrates Business Rules and Predictive Analytics techniques. It allows business users to define training rules specific for their datasets. Then Rule Learner will build the proper training set, execute a machine learning algorithm and generate classification rules in OpenRules and PMML formats. Our multi-year IRS experience “Machine Learning Models Research and Development with NRP and LMSB Data” has proven practicality of this approach.
Our recent results in the area of conflicting rules and defeasible logic are described here. If I can find time, I will try to prepare a paper for the above RuleML event.