Comparison of interestingness measures applied to textual taxonomies matching

This paper presents an experimental comparison of Interestingness Measures (IMs), in the context of an approach designed for matching textual taxonomies. This extensional and asymmetric approach makes use of association rule model for matching entities issued from two textual hierarchies. We select 6 IMs and we perform two experiments on a benchmark composed of two textual taxonomies and a set of reference matching relations between the concepts of the two structures. The first test concerns a comparison of matching accuracy with each of the selected measures. In the second experiment, we compare how each IM evaluates reference relations by studying their values distributions. Results show that the implication intensity delivers the best results.