• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Book chapter

Adaptive Multi-model Approaches to Pattern Set Mining

P. 1-8.
Kuznetsov S., Makhalova T., Napoli A.

Pattern Mining (PM) has an important place in Data Mining and Knowledge Discovery and has many applications in a wide variety of domains. To date, a lot of different approaches to PM have been pro­posed. However, new methods continue to appear. Some of the reasons for that are the following: (i) there is no gold standard for evaluating the quality of PM approaches, (ii) the results of existing approaches are unsatisfactory. But what is wrong with them? In this paper, we adopt the best practices of building supervised models to PM. We propose a PM method Keeplt Simple that combines the technique of supervised learning and the state-of-the-art of modern PM. We show in experi­ments that the proposed approach allows for obtaining small sets of interesting and non-redundant patterns.