Abstract
We propose the ontological approach to Data Mining that is based on: (1) the analysis of subject domain ontology, (2) information in data that are interpretable in terms of ontology, and (3) interpretability of Data Mining methods and their results in ontology. Respectively concepts of Data Ontology and Data Mining Method Ontology are introduced. These concepts lead us to a new Data Mining approach—Ontological Data Mining (ODM). ODM uses the information extracted from data which is interpretable in the subject domain ontology instead of raw data. Next we present the theoretical and practical advantages of this approach and the Discovery system that implements this approach. The value ofODMis demonstrated by solutions of the tasks from the areas of financial forecasting, bioinformatics and medicine.
Original language | English |
---|---|
Pages (from-to) | 277-292 |
Number of pages | 16 |
Journal | Studies in Computational Intelligence |
Volume | 683 |
DOIs | |
Publication status | Published - 1 Jan 2017 |