The present task describes the participation of the JU_NITM team in IJCNLP-2017 Shared Task 5: "Multi-choice Question Answering in Examinations". One of the main aims of this shared task is to choose the correct option for each of the multi-choice questions. We represent each of the questions and its corresponding answer in vector space and find the cosine similarity between two vectors. Our proposed model also includes supervised classification techniques to find the correct answer. Our system was only developed for the English language, and it also obtains an accuracy of 40.07% for the test dataset and 40.06% for validation dataset, respectively.
展开▼