Automatic Evaluation of Quality of an Explanatory Dictionary by Comparison of Word Senses
- Cite this paper as:
- Gelbukh A., Sidorov G., Han S., Chanona-Hernandez L. (2004) Automatic Evaluation of Quality of an Explanatory Dictionary by Comparison of Word Senses. In: Broy M., Zamulin A.V. (eds) Perspectives of System Informatics. PSI 2003. Lecture Notes in Computer Science, vol 2890. Springer, Berlin, Heidelberg
Words in the explanatory dictionary have different meanings (senses) described using natural language definitions. If the definitions of two senses of the same word are too similar, it is difficult to grasp the difference and thus it is difficult to judge which of the two senses is intended in a particular contexts, especially when such a decision is to be made automatically as in the task of automatic word sense disambiguation. We suggest a method of formal evaluation of this aspect of quality of an explanatory dictionary by calculating the similarity of different senses of the same word. We calculate the similarity between two given senses as the relative number of equal or synonymous words in their definitions. In addition to the general assessment of the dictionary, the individual suspicious definitions are reported for possible improvement. In our experiments we used the Anaya explanatory dictionary of Spanish. Our experiments show that there are about 10% of substantially similar definitions in this dictionary, which indicates rather low quality.
Unable to display preview. Download preview PDF.