Data mining based pervasive system design for Intensive Care Unit

Intensive Care Unit is extremely useful for providing health care to critically ill patients for their speedy recovery. An Intensive Care Unit (ICU) consists of multiple life saving machines which are continuously recording vital organs of patients and generating vast data at very high frequency. A context aware based real time monitoring system is extremely useful in the modern era of bio-medical engineering which could able to help medical expertise for efficient decision making. Considering multiple vital factors in a real-life scenario an intelligent decision support system could be developed and work more efficiently if associated with Pervasive Data Mining techniques. The approach presented in this research work illustrates the Pervasive Data Mining for Intensive care unit which utilizes “Adaptive Direct Acyclic Graph based Support Vector Machine” (ADAGSVM) to classify the health conditions of elderly patients. An Adult Prognostic Index (API) scoring system has been utilized which could be calculated on the basis of six important vital signs. Based on the value of Adult Prognostic Index, clinical rules may be identified to initiate alarms for monitoring of patient health conditions.