Increased predictive value of parameters by fuzzy logic‐based multiparameter analysis

Abstract

Background: A recent study on postoperative effusions and edema was used to demonstrate the potential of fuzzy techniques in multiparameter data analysis. In this study, more than 50 parameters of 75 patients were collected and examined for correlations between some of the parameters and the later development of complications.Methods: We employed a rule-based fuzzy-logic system in order to combine the diagnostic values of single parameters. The advantage of fuzzy sets is that they substitute sharp cut-off values with a smooth transition from one property to another. Therefore, there is no decision of "either-or" but rather a graded assessment of "more or less", which is often more suitable for a problem.Results: The fuzzy combination of parameters led to a large increase of sensitivity and specificity when compared with the best single parameter. This increase was achieved by taking a close look at the parameters. A newly created parameter, relative weight, turned out to be very powerful.Conclusions: Fuzzy techniques can increase the discriminating power of classical statistical tools. In addition, results obtained by fuzzy analysis are highly interpretable. A combination of the CLASSIF1 algorithm for the identification of the most relevant parameters, followed by fuzzy analysis, represents a powerful tool for the handling of large amounts of multiparameter data. Cytometry Part B (Clin. Cytometry) 53B:75-77, 2003.

Cytomics in predictive medicine

“…Currently, this information is frequently extracted in a fragmentary way by computer‐assisted identification and characterization of a few cell populations or gene clusters of interest. Alternatively, all of the available information can be screened exhaustively by multiparametric clustering (17, 18), data mining, or other procedures (19) for diagnostic or prognostic information‐using hypothesis‐driven analysis strategies. These frequently require mathematical, statistical, or other assumptions that may unintentionally bias the results.…”