A Collaborative Relevance Feedback Approach to Task-driven Recommendation
I‐Chin Wu, Duen‐Ren Liu · Journal of the Association for Information Systems · 2003
Managing knowledge is an important issue for organizations to gain sustainable advantage in today's competitive environments.Nevertheless, the operations of enterprises are mainly planned around task execution.Accordingly, it is important to provide knowledge support for task execution by recommending task relevant knowledge to fit the information needs of task execution.This paper contributes to the stream of knowledge management system by proposing a novel task-driven recommender system to take the merits of information retrieval and cooperative work.A task-oriented information repository based on fuzzy classification is deployed to support task-driven recommendation.Additionally, the system uses task specification to model the key contents of executing-task and to further facilitate the process of identifying relevant information needs of task execution.A collaborative relevance feedback approach is proposed to adjust task specification with the aid of cooperative workers and domain experts.The proposed system enhanced with the adaptation capability to adjust task specification makes the recommendation of relevant information more flexible and adaptable to dynamically-changing working environments.