Learning Decision Trees Using the Area Under the ROC Curve
ICML '02 Proceedings of the Nineteenth International Conference on Machine Learning
| Hi-index | 0.00 |
This paper focuses on the extraction of lexical knowledge from text by exploiting the "glosses" of words, i.e. spontaneous descriptions identifiable by lexical markers and specific morpho-syntactic patterns. This information offers interesting knowledge in order to build dictionaries. In this study based on the RESENS project, we compare two methods to extract this linguistic information using local grammars and/or web-mining approaches. Experiments have been conducted on real data in French.