[ English | Japanese ]
Keio University, Graduate School of Media and Governance
Back to Index Page
Research on use of social contexts for information retrieval on
The object of this research is to enable the use social contexts for information retrieval in the Internet, thereby realizing needs which traditional methods could not satisfy.Internet information retrieval differs from traditional information retrieval, such as within a closed organization or in an library, in that it is difficult to match the users search needs with the context of the area of search and further, that the reliability of the obtained information is hard to evaluate. This research proposes two methods which solve these issues by incorporating into the information retrieval process, the cognitive patterns shared by a social grouping, called "Social Context".The three types of social context referred to in this research are "Semantic Relativity within a Specialized Field", "Trust of an Authoritative Organization" and "Trust of Human Networks". These map directly to three types of information structures often seen on the Internet, "Shared Database Structure", "Hie!
rarchical Structure" and "Self-Originating Network Structure".The first method proposed in this research is for creating and using meta-data spaces for semantic searches. In semantic searches, the meta-data space is used as a reference for queries and measuring the contextual relativity of the area of search. In the proposed method, by making use of the document set of a specific discipline along with a thesaurus, a meta-data space is created, expressing the semantic context shared by specialists of the given discipline for specialized terms. Furthermore, a method for utilizing the created meta-data space to support the use of specialized terms in the expression of queries is also proposed.In the second method, the mechanisms for digital certificate authentication is combined with that of database searches. In this method, PKI is used as the mechanism to evaluated the reliability of information found on the Internet. In this method the hierarchicallystructured PKIX is used as!
the PKI which expresses the social context of trust in an authoritative organization, while a new certificate validation model called PGN is proposed as the PKI to express trust in human networks. By merging database search functions with the two mechanisms that validate the certificates found within the databases, PKIX and PGN, an information retrieval method for evaluating the reliability of the information found with in the database is realized.
To obtain the whole paper, please contact;
[ English | Japanese ]
Back to Project Home Page