On the use of multi-objective evolutionary algorithms for the induction of fuzzy classification rule systems.

Extracting comprehensible and general classifiers from data in the form of rule systems is an important task in many problem domains. This study investigates the utility of a multi-objective evolutionary algorithm (MOEA) for this task. Multi-objective evolutionary algorithms are capable of finding several trade-off solutions between different objectives in a single run. In the context of the present study, the objectives to be optimised are the complexity of the rule systems, and their fit to the data. Complex rule systems are required to fit the data well. However, overly complex rule systems often generalise poorly on new data. In addition they tend to be incomprehensible. It is, therefore, important to obtain trade-off solutions that achieve the best possible fit to the data with the lowest possible complexity. The rule systems produced by the proposed multi-objective evolutionary algorithm are compared with those produced by several other existing approaches for a number of benchmark datasets. It is shown that the algorithm produces less complex classifiers that perform well on unseen data.

Early identification of sepsis and mortality risks through simple, rapid clot-waveform analysis. Implications of lipoprotein-complexed C reactive protein formation.

OBJECTIVE: To determine if the rapid waveform profile of the activated partial thromboplastin time (aPTT) assay, which detects lipoprotein-complexed C reactive protein (LCCRP) formation, predicts sepsis and mortality in critically ill patients. DESIGN: Observational, cohort study. SETTING: General intensive therapy unit (ITU) of a tertiary care hospital. PATIENTS AND PARTICIPANTS: A total of 1187 consecutive patients admitted to the ITU. INTERVENTION: Activated partial thromboplastin time transmittance waveform analysis was performed within the first hour of admission to the ITU. The degree of change causing a biphasic waveform was quantified through the drop in light transmittance level. MEASUREMENTS AND RESULTS: Three hundred forty-six patients had a biphasic waveform on admission to the ITU with a mortality rate of 44% compared with 26% for those with normal waveforms. Logistic regression models showed direct correlation between the likelihood for sepsis and in-patient mortality with increasing waveform abnormalities. The mortality fraction was 0.3 with normal waveforms versus 0.6 when the light transmittance decreased by 30%. The odds ratio (OR) for mortality and sepsis were 4.5 and 11, respectively, from the most abnormal to normal aPTT waveforms. These were comparable with APACHE II scores and superior to those estimated by CRP for mortality (OR 2.3) / sepsis (OR 6.4) prediction. CONCLUSION: Waveform analysis within the first hour of ITU admission is a single, simple and rapid method of identifying the risks of mortality and sepsis. Its measure of LCCRP formation shows superior prediction over CRP alone and it warrants further assessment as a tool to triage and target prompt, appropriate treatment in the ITU.