Towards Automatic Classification of Description of Analogies into SAPPhIRE Constructs

The motivation behind this research is to aid designers in retrieving a large number of analogies relevant to the design problem under consideration. The objective of the work reported here is to automatically classify natural-language descriptions of analogies into constructs of SAPPhIRE model of causality-a common causal language that was developed for describing the functioning of natural and artificial systems. This should provide the benefit of utilizing the enormous data available on the Internet, while also providing focused search results. Using supervised classification approach, the accuracy (cross-validation) achieved in classification of natural-language descriptions of analogies into the constructs of SAPPhIRE model was 0.70 +/- 0.08.