Extending data mining techniques for frequent pattern discovery: Trees, low-entropy sets, and crossmining