Search Intenion Network for Personalized Query Auto-Completion in E-Commerce

Query Auto-Completion(QAC), as an important part of the modern Generally, in search engines, traditional QAC system follows a search engine, plays a key role in complementing user queries two-stage method: matching and ranking. In the matching phase, and helping them refine their search intentions. Today's QAC systems a sufficient number of candidate queries matching the prefix are in real-world scenarios face two major challenges:1)intention recalled from the history log. In the ranking stage,the candidate equivocality(IE): during the user's typing process, the prefix often historical frequency features[3, 24, 45] and semantic features[22, contains a combination of characters and subwords, which makes 32, 44] are used to obtain the final list ranking order. Finally, due to the current intention ambiguous and difficult to model.2)intention the limitation of display space, several top ranked candidates will transfer (IT):previous works make personalized recommendations be provided to users.