An Unsupervised Keyword Query Segments Representation Approach Using Word embeddings for XML retrieval
XML keyword Search systems retrieves high quality and most relevant results when they are able to identify the important phrases in the keyword query which need to be kept together for quality results. One way to achieve this is for the user to be explicit about the phrases by adding quotes around the segments of the keywords query to indicate phrases. But this is not the pattern in the real-world search logs. Users expect the keyword search system to infer and understand these phrases. In this research, we proposed an unsupervised keyword query segments representation approach for xml retrieval. Our approach outperforms existing approaches in terms of precision.
Download Paper