Mining Clinical Data with a Temporal Dimension: A Case Study

摘要: Clinical databases store large amounts of information about patients and their medical conditions. Data mining techniques can extract relationships patterns holding in this wealth data, thus be helpful understanding the progression diseases efficacy associated therapies. A typical structure data is a sequence observations clinical parameters taken at different time moments. In kind contexts, temporal dimension fundamental variable that should account process returned as part extracted knowledge. Therefore, classical well established framework sequential pattern not enough, because it only focuses on sequentiality events, without extracting elapsing between two particular events. Time-annotated sequences (IAS), novel paradigm solves problem. Recently defined our laboratory together with an efficient algorithm for them, IAS are where each transition events annotated found frequent data. paper we report real-world case study, which applied to regarding set follow-up liver transplantation. The aim analysis assessing effectiveness extracorporeal photopheresis (ECP) therapy prevent rejection solid organ For patient, biochemical variables recorded moments after show values interleukins other specific dates, from possible physician assess ECP therapy. We believe study does interestingness context but, more ambitiously, suggests general methodology mining, whenever important problem analysis.