Exploring the Query-Flow Graph with a Mixture Model for Query Recommendation

Query recommendation has been recognized as an important tool that helps users in their information seeking activities. Many existing approaches leveraged the rich Web query logs to generate query recommendations. Recently, the query-flow graph, an aggregated representation of session information in query logs, has shown its utility in query recommendation. However, there are two major problems in directly using query-flow graph for recommendation. On one hand, due to the sparsity of the graph, one may not well handle the recommendation for many dangling queries in the graph. On the other hand, without addressing the ambiguous intents in such an aggregated graph, one may generate recommendations either with multiple intents mixed together which are difficult to consume, or dominated by certain intent which cannot satisfy different user needs. In this paper, we propose to explore the queryflow graph with a mixture model for better query recommendation. Specifically, we propose a novel mixture model that describes the generation of the query-flow graph. With this model, we can identify the hidden intents of queries from the graph. We then apply an intent-biased random walk over the graph for query recommendation. In this way, we can well resolve the above two problems. Some primary experiments on real query logs show the effectiveness of our approaches as compared with baseline methods.