Nowadays, human-computer information retrieval (HCIR) is one of the most researched approaches, which applies human computer interaction (HCI) in information retrieval (IR), especially for text retrieval. The involvement of human activity (i.e. intelligence) in the process of retrieval is an interesting and challenging aspect of HCIR, thus it takes an interactive character with the purpose to improve user's needs. A number of researchers and research groups have conducted different experiments in this area. Innovative techniques for determining relevant documents and navigating in a large and complex set of full document results have been proposed. A novel approach based on focused retrieval for interactive identification of relevant document's parts has brought many improvements in this field. A special emphasis was placed on retrieving passages and XML elements from structured documents. Numerous techniques for focused retrieval based on snippets, facets and relevance feedback have been presented recently, bringing advantages and a more realistic scenario for text retrieval, but also a lot of open issues that need to be explored. This paper surveys the fundamental characteristics of the HCIR, as well as the focused HCIR, such as using passages and XML elements as retrievable units from structured or pure textual documents, thus providing a basis for development of novel HCIR approaches for interactive text retrieval, as our main goal in future research.
HCIR, information retrieval, passage, XML element, evaluation