In this paper we propose an hybrid system of Arabic words disambiguation. To achieve this goal we use the methods employed in the domain of information retrieval: Latent semantic ...
This paper presents two fast and robust algorithms for word skew and slant corrections based on Radon transform. For the skew correction, we maximize a global measure which is deļ...
Jian-xiong Dong, Dominique Ponson, Adam Krzyzak, C...
Existing word similarity measures are not robust to data sparseness since they rely only on the point estimation of words' context profiles obtained from a limited amount of ...
Jun'ichi Kazama, Stijn De Saeger, Kow Kuroda, Masa...
Abstract. In many applications of natural language processing (NLP) it is necessary to determine the likelihood of a given word combination. For example, a speech recognizer may ne...
The annotation of words and phrases by ontology concepts is extremely helpful for semantic interpretation. However many ontologies, e.g. WordNet, are too ļ¬ne-grained and even hu...