MIKIDQD: Machine Intelligence and Knowledge Infused Duplicate Question Detection

Quora is a flourishing online question answer forum where millions of users post questions and answer questions in the subjects they have knowledge. A plethora of questions are posted by users and some of them can be similarly worded. This would make it tedious for the users as answering similarly worded questions can be time consuming. Hence, there is a need to detect similarly worded questions or duplicate questions. This research work proposes a MIKIDQD framework for detecting duplicate questions in the Quora forum by using the Quora dataset. The knowledge-centric, ontologybased and semantically-driven strategy incorporated in the proposed model surpasses the other baseline models in terms of the performance measures. The proposed model yields the highest accuracy value of 98.07% and the least False Negative Rate of 0.01 indicating the best performance.