Paula Vaz Lobo


2010

pdf bib
Named Entity Recognition in Questions: Towards a Golden Collection
Ana Cristina Mendes | Luísa Coheur | Paula Vaz Lobo
Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)

Named Entity Recognition (NER) plays a relevant role in several Natural Language Processing tasks. Question-Answering (QA) is an example of such, since answers are frequently named entities in agreement with the semantic category expected by a given question. In this context, the recognition of named entities is usually applied in free text data. NER in natural language questions can also aid QA and, thus, should not be disregarded. Nevertheless, it has not yet been given the necessary importance. In this paper, we approach the identification and classification of named entities in natural language questions. We hypothesize that NER results can benefit with the inclusion of previously labeled questions in the training corpus. We present a broad study addressing that hypothesis, focusing on the balance to be achieved between the amount of free text and questions in order to build a suitable training corpus. This work also contributes by providing a set of nearly 5,500 annotated questions with their named entities, freely available for research purposes.

pdf bib
Fairy Tale Corpus Organization Using Latent Semantic Mapping and an Item-to-item Top-n Recommendation Algorithm
Paula Vaz Lobo | David Martins de Matos
Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)

In this paper we present a fairy tale corpus that was semantically organized and tagged. The proposed method uses latent semantic mapping to represent the stories and a top-n item-to-item recommendation algorithm to define clusters of similar stories. Each story can be placed in more than one cluster and stories in the same cluster are related to the same concepts. The results were manually evaluated regarding the groupings as perceived by human judges. The evaluation resulted in a precision of 0.81, a recall of 0.69, and an f-measure of 0.75 when using tf*idf for word frequency. Our method is topic- and language-independent, and, contrary to traditional clustering methods, automatically defines the number of clusters based on the set of documents. This method can be used as a setup for traditional clustering or classification. The resulting corpus will be used for recommendation purposes, although it can also be used for emotion extraction, semantic role extraction, meaning extraction, text classification, among others.