Driven Rule Mining and Representation of Temporal Pattern in Physiological Sensor Data

Nilam Divekar, S. V. Todakari

Abstract


Mining and representation of subjective examples is a developing field in sensor information examination. This paper influences from standard mining strategies to extricate and speak to worldly connection of prototypical examples in clinical information streams. The methodology is completely information driven, where the fleeting principles are mined from physiological time arrangement, for example, heart rate, breath rate, and blood weight. To accept the tenets, a novel closeness technique is presented, that analyzes the similitude between guideline sets. An extra part of the proposed approach has been to use characteristic dialect era methods to speak to the transient relations between examples. In this study, the sensor information in the MIMIC online database was utilized for assessment, in which the mined fleeting rules as they identify with different clinical conditions (respiratory disappointment, angina, sepsis etc) were made express as a printed representation. Moreover, it was demonstrated that the removed tenet set for a specific clinical condition was particular from other clinical condition


Keywords: Health informatics, Data-driven modelling, pattern abstraction, physiological sensor data, sensor data analysis, temporal rule mining.


Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v7i4.2694

Refbacks

  • There are currently no refbacks.




Copyright (c) 2016 International Journal of Advanced Research in Computer Science