FreeSpan: frequent pattern-projected sequential pattern mining
Proceedings of the sixth ACM SIGKDD international conference on Knowledge discovery and data mining
ICDE '95 Proceedings of the Eleventh International Conference on Data Engineering
Frequent pattern mining: current status and future directions
Data Mining and Knowledge Discovery
Efficient mining of frequent episodes from complex sequences
Information Systems
Temporal Data Mining
Universal glucose models for predicting subcutaneous glucose concentration in humans
IEEE Transactions on Information Technology in Biomedicine
Analysis of medical pathways by means of frequent closed sequences
KES'10 Proceedings of the 14th international conference on Knowledge-based and intelligent information and engineering systems: Part III
| Hi-index | 0.00 |
The disease of diabetes mellitus has spread in recent years across the world, and has thus become an even more important medical problem. Despite numerous solutions already proposed, the problem of management of glucose concentration in the blood of a diabetic patient still remains as a challenge and raises interest among researchers. The data-driven models of glucose-insulin interaction are one of the recent directions of research. In particular, a data-driven model can be constructed using the idea of sequential patterns as the knowledge representation method. In this paper a new hierarchical, template-based approach for mining sequential patterns is proposed. The paper proposes also to use functional abstractions for the representation and mining of clinical data. Due to the experts knowledge involved in the construction of functional abstractions and sequential templates, the discovered underlying template-based patters can be easily interpreted by physicians and are able to provide recommendations of medical therapy. The proposed methodology was validated by experiments using real clinical data of juvenile diabetes.