Abstract
Thousands of complex natural language questions are submitted to communityquestion answering websites on a daily basis, rendering them as one of the mostimportant information sources these days. However, oftentimes submittedquestions are unclear and cannot be answered without further clarificationquestions by expert community members. This study is the first to investigatethe complex task of classifying a question as clear or unclear, i.e., if itrequires further clarification. We construct a novel dataset and propose aclassification approach that is based on the notion of similar questions. Thisapproach is compared to state-of-the-art text classification baselines. Ourmain finding is that the similar questions approach is a viable alternativethat can be used as a stepping stone towards the development of supportive userinterfaces for question formulation.