Automatic Multiple Choice Question Generation System for Semantic Attributes Using String Similarity Measures

Abstract

This research introduces an automatic multiple choice question generation system to evaluate the understanding of the semantic role labels and named entities in a text. The system provided selects the informative sentence and the keyword to be asked based on the semantic labels and named entities that exist in the sentence, the distractors are chosen based on a similarity measure between sentences in the data set. The system is tested using a set of sentences extracted from the TREC 2007 dataset for question answering. From the experimental results, it can be induced that the semantic role labeling and named entity recognition approaches could be used as   a good keyword selection mechanism. The second conclusion is that the string similarity measures proved to be a very good approach that can used in generating the distractors for an automatic multiple choice question. Also, combining the similarity measures of different algorithms would lead to generate a good distractors.