DIMSUM: an expert system for multiexponential model discrimination

Abstract
DIMSUM is a highly automated, rule-based expert system designed to fit multiexponential models of increasing dimension to time series data, followed by selection of the best candidate model based on a user-modifiable and weighted decision tree of statistical criteria for model discrimination. The major features of DIMSUM are 1) an interactive and friendly user interface; 2) options for incorporating prior information about the parameters, the data, and/or the system from which the data were collected, in the form of equality and inequality constraints; 3) a built-in algorithm for automatically obtaining starting values for parameter estimation; 4) a robust weighted least-squares parameter estimation algorithm operating in an adaptive, user-adjustable search space; 5) comprehensive statistical results comparing different order candidate models fitted to the data; and 6) a novel, user-modifiable (learning) rule-based advisory subsystem providing an “expert's” interpretation of these statistical results and an explanation of all advice.

This publication has 4 references indexed in Scilit: