On the effectiveness of feature set augmentation using clusters of word embeddings
This work addresses feature engineering for NLP practitioners, but it is incremental as it confirms existing empirical findings through systematic evaluation.
The paper tackled the problem of understanding the role of word cluster features in NLP tasks by systematically evaluating their effect on named entity segmentation, classification, sentiment classification, and quantification, finding that cluster membership features improve performance.
Word clusters have been empirically shown to offer important performance improvements on various tasks. Despite their importance, their incorporation in the standard pipeline of feature engineering relies more on a trial-and-error procedure where one evaluates several hyper-parameters, like the number of clusters to be used. In order to better understand the role of such features we systematically evaluate their effect on four tasks, those of named entity segmentation and classification as well as, those of five-point sentiment classification and quantification. Our results strongly suggest that cluster membership features improve the performance.