NUTKS at NTCIR-12 MobileClick2: iUnit Ranking Subtask Using Topic Model

In this paper, NUTKS (Nagaoka University of Technology, Knowledge Systems Laboratory) reports the results of our participation at the NTCIR-12 MobileClick task, iUnit Ranking subtask. The authors have ranked iUnit by using similarity of iUnit word distribution based on LDA topics. Our system has recorded Q-measure score as 0.7392 and nDCG@20 score as 0.6334. In baseline system, the recorded Qmeasure score is 0.7411[1]. Therefore, there is no significance difference between the baseline system and our proposed approach. According to these results, the search intent is considered as topics. However, it cannot output result of reflecting search intent diversity only using topic similarity. it is not strong enough to reflect the diversity of search intent by using only the topic similarity.