Automatic Detection of Search Tactics in Collaborative Exploratory Web Search Process

Information seeking process is an important research topic in information seeking behavior. Collaborative information seeking (CIS) has attracted many researchers’ attention in recent years, but the investigation of CIS process is still rare. Investigations on search processes can either be macro-level or micro-level. The macro-level investigation focuses on establishing theoretical models while micro-level investigation focuses on identifying descriptive categories such as user action or search tactics. In this paper, we proposed an automatic technique and explicitly model the latent search tactics using a Hidden Markov Model. HMM results show that the identified search tactics transition patterns in individual information seeking process are consistent with Marchionini’s information seeking process model. Then, we applied the HMM in CIS and found different patterns of search tactics compared to the individual search. With the advantages of showing the connections between search tactics and search actions, and the transitions among search tactics, we argue that Hidden Markov Model is a useful tool to investigate information seeking process, or at least it provides a feasible method to analyze large scale dataset.