The Treatment of Time in a Case-Based Analysis of Experimental Medical Studies

摘要

Case-based approaches are employed within a multitude of application areas one of which is the prediction of dynamic behaviour. Given a situation the possible development after a time span shall be determined. If only a small set of heterogenously structured cases describing observations at a variety of time points is given to start with, as it is the case when experimental medical studies shall be analysed, it becomes necessary to analyse and evaluate the temporal horizon from case to case differently, and treat time as a first class variable. This is the strategy OASES (Our Approach to Simulation based on Experimental Studies) employs. For this purpose, OASES utilises knowledge implicit in cases for matching and adaptation. Whether different time points do match in the current situation or how different time points of observation might effect the development, is decided based on cases which time is an explicit part of. Keywords: Case-Based Reasoning; Case-Based Similarity Ranking; Case-Based Adaptation; Prediction; Experimental Studies