Instance pruning by filtering uninformative words: an information extraction case study

In this paper we present a novel instance pruning technique for Information Extraction (IE). In particular, our technique filters out uninformative words from texts on the basis of the assumption that very frequent words in the language do not provide any specific information about the text in which they appear, therefore their expectation of being (part of) relevant entities is very low. The experiments on two benchmark datasets show that the computation time can be significantly reduced without any significant decrease in the prediction accuracy. We also report an improvement in accuracy for one task.