Evaluation of the super-modified simplex for use in chemical pattern recognition

Multivariate rule building expert system

“…A function optimization approach is used to find the transformation vector, T, that minimizes the entropy of classification H(C\A). Optimization has been used to generate linear discriminants in previous studies which have been reported to be superior in performance to those obtained from the LLM method (18)(19)(20). A modified simplex was used for the classification results reported and is described in detail in ref 20.…”

“…Optimization has been used to generate linear discriminants in previous studies which have been reported to be superior in performance to those obtained from the LLM method (18)(19)(20). A modified simplex was used for the classification results reported and is described in detail in ref 20. The simplex differs from the one cited in that it operates on the original u-dimensional data space and the v + 1 component is determined from the algorithm to be given later.…”