Cluster Analysis in Clinical Chemistry: A Model

Provides a new statistical diagnostic model. The authors discuss the theoretical background of diagnosing in detail, the most important analytical methods being revealed in their discussion of cluster analysis. Problems associated with commonly-used cluster analytic algorithms are addressed carefully. The authors go on to describe a strategy algorithm in a general setting, then apply their cluster analytic strategy algorithm to a diagnostic problem. The example used employs a real data set on thyroid diseases. The application of the diagnostic model to this data set is examined in detail and the clinical usefulness of the model is confirmed by means of a second data set. The model is adaptable enough to find application to other diagnostic problems. Includes a Fortran computer program.