Prediction of learner’s appropriate online community of practice in question and answering website: similarity in interaction, interest, prior knowledge

Abstract

ABSTRACT The development of information technology and social networks has created new opportunities to access lifelong learning in the form of informal learning. In an informal learning environment, learning takes place via Communities of Practice (CoP). The learning success factors in online CoPs are learners’ similarity in learning interests and learners’ ability to interact with each other in the community. In this paper, a method based on Bayes probability and modified collaborative filtering is proposed to predict next learner topics and suggests appropriate learning questions. In the recommended approach, the learner’s CoP predictions are made based on learners’ behavior in the learning environment and their prior knowledge. The proposed method is evaluated using the dataset of Stack Overflow as a professional Question Answering website. Results show: 1) The proposed method with cosine similarity function, has better performance than traditional Collaborative Filtering, in the prediction of question related to CoP based on co-learner and mentor roles. 2) Using the proposed method can increase the effective appropriate interaction ratio in the learning environment.