A New Approach for Finding and Ranking Question – Answering Pairs in Community Question Answering

Objective: To help users receive immediate answers when they submit an input question to the community Question Answering (cQA) websites. Methods: To find and ranking question - answering pairs in cQA related to the input question, in this paper we investigate various aspects of understanding questions as well as answers to extract different features kinds from them. Then all these obtained features are combined into a machine learning based framework for getting similarity existing question - answering pairs as well as for ranking these pairs. Finding: Besides the traditional features such as bag of n-grams we will use more efficient aspects that include word embeddings, question categories and keywords extraction. We will use a word representation model for generating word embeddings, a question categorization module for determining the category for an input question and keywords extraction module to extract important words (phrases) in each question - answering pair. We also use the SVM and MLP classifiers to generate the predicted scores and then use these scores to the ranking question - answering pairs. The experiment shows obtained results; the MLP classifier gives higher results than the SVM in both Accuracy and MAP measures. Improvements: Our model has achieved better results in comparison with the previous studies on the same dataset. The Accuracy and MAPmeasures increased by 2.86% and 1.65%, respectively, compared with the previous best result.