Autonomous interpretation of statistical analysis for engineering decision making
Availability Unrestricted Thesis/Dissertation
Abstract
Existing statistical software fails to explain the meaning of their output. Practicing engineers, in areas of application where statistics is heavily used, have to deal not only with the nuances of different statistical packages but also with learning what the results from their analyses mean. An Architecture and a prototype for a knowledge-based statistical output interpreter has been designed. CLIENS integrates different heterogeneous components and it uses object oriented design principles to enable modularity and package independence. Outputs from several statistical packages were analyzed to discover common patterns. A heuristic was developed to search automatically for these patterns in the output files. An in-depth study of a small set of statistical techniques resulted in the derivation of descriptive and inferential knowledge, which is used by CLIENS to interpret statistical outputs. Experimentation with the prototype indicates that autonomous interpretation is feasible and that package independence is achievable. However, issues pertaining to natural language need to be resolved before a commercial CLIENS exists.
Rights Statement
In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/
This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).