Identifying targets for syntactic simplification
Julie Medero, Mari Ostendorf · 2011
Most work on automatic text simplification considers lexical difficulty separate from syntactic simplification.In this study, we use both factors together to predict a variety of sentence changes, including the standard problems of splitting and shortening, as well as expanding to define difficult words that are important to the topic.We leverage a variety of lexical and parse features, as well as a score of the relatedness of a sentence to the topic of its document.