Ampliación automática de corpus mediante la colaboración de varios etiquetadores

The availability of extense tagged data corpus is an essential aspect in many NLP tasks. The effort required for tagging manually this large number of phrases has encouraged many researchers like us to create automatic applications for this issue. Our approach represents a completely automatic method (optionally applying a minimum effort) for enlarging an already existing corpus, so it acquires the desired number of tagged phrases. The extra content of the corpus will be obtained from any knowledge source like the web, from where we extract untagged sentences to be analyzed. Considering the initial small corpus as the seed, our method makes it evolve until it reaches the goal size. The process is based on several taggers using the co-training technique, achieving the results after a number of iterations and applying the stacking scheme for deciding which new tagged sentences must be incorporated to the new corpus.