Factorial Design Expert System
An automated expert system that uses split-run and factorial design methods to determine which factors are most important in an experiment. The expert system is architected into Design, Execute and Evaluate phases, to assist a user in developing a Factorial Design experiment in which one, two or three factors are tested simultaneously. In a preferred embodiment, a database infrastructure and web client, browser-based methodology functions as the expert system (a “wizard”) to design experiments, build control groups and evaluate results, all with the goal of discovering what values for which factors will yield the optimum response from subjects.
1 . A method for design of a test, the method comprising the steps of:
designing the test, using an interactive test design process executed in an integrated data processing environment, the test including at least a specification of test type, at least one test group of consumers against which the test is to be run, and multiple factors to be tested simultaneously as independent design variables;
to executing the test within an online expert system to determine test results, the test results including responses by the at least one test group of consumers to the multiple factors; and
evaluating the test results by determining at least one of a main effect or interaction effect of the multiple factors.