Knowledge discovery within the health informatics domain provides healthcare professionals with the ability to further inform their decisions regarding patient treatment. However discovering patterns within this domain has proven difficult for data scientists. One of the main challenges is finding meaningful patterns that can be used as effective support alongside specialist knowledge. This paper demonstrates the value behind extracting temporal patterns from live datasets. In particular it shows how fuzzy logic and divisive hierarchical clustering can be used to extract frequent sequential patterns with time-intervals from a series of breast cancer diagnoses. A discussion regarding the link between time-intervals and cancer treatment is explored through the use of multi-dimensional sequential patterns mining with fuzzy time-intervals.
|Title of host publication||Thirty-fifth SGAI International Conference on Artificial Intelligence, 15-17 December 2015, Cambridge.|
|Number of pages||6|
|Publication status||Published - 1 Dec 2015|