Automated Suggestions for Miscollocations

One of the most common and persistent error types in second language writing is collocation errors, such as learn knowledge instead of gain or acquire knowledge, or make damage rather than cause damage. In this work-inprogress report, we propose a probabilistic model for suggesting corrections to lexical collocation errors. The probabilistic model incorporates three features: word association strength (MI), semantic similarity (via WordNet) and the notion of shared collocations (or intercollocability). The results suggest that the combination of all three features outperforms any single feature or any combination of two features. 1 Collocation in Language Learning The importance and difficulty of collocations for second language users has been widely acknowledged and various sources of the difficulty put forth (Granger 1998, Nesselhauf 2004, Howarth 1998, Liu 2002, inter alia). Liu’s study of a 4million-word learner corpus reveals that verb-noun (VN) miscollocations make up the bulk of the lexical collocation errors in learners’ essays. Our study focuses, therefore, on VN miscollocation correction. 2 Error Detection and Correction in NLP Error detection and correction have been two major issues in NLP research in the past decade. Projects involving learner corpora in analyzing and categorizing learner errors include NICT Japanese Learners of English (JLE), the Chinese Learners of English Corpus (Gamon et al., 2008) and English Taiwan Learner Corpus (or TLC) (Wible et al., 2003). Studies that focus on providing automatic correction, however, mainly deal with errors that derive from closed-class words, such as articles (Han et al., 2004) and prepositions (Chodorow et al., 2007). One goal of this work-in-progress is to address the less studied issue of open class lexical errors, specifically lexical collocation errors.