HYKE LogoSemantic Search


Home
Introduction
State of the art
An overview of SemSearch
The Google-like query interface
Making sense of the user query
Translating the user query into formal queries
Implementation and experimental evaluation
Conclusions and future work
1st Workshop on Semantic Search
2nd Workshop on Semantic Search
3rd Workshop on Semantic Search
4th Workshop on Semantic Search
 

The Google-like query interface


The query interface of SemSearch extends traditional keyword search languages
by allowing the explicit specification of i) the queried subject and ii) the combi-
nation of keywords. By the term subject, the user can explicitly tell the search
engine the expected type of the search results. For example, when the user spec-
ifies the keyword “news” as the subject keyword, he or she expects the search
results to be the instances of the class entity that matches against the keyword
“news”. By allowing the specification of how to combine multiple keywords, the
search engine provides end users a simple way of expressing complex queries.
SemSearch uses three heuristic operators to support the specification of user
queries. The first one is the symbol “:”, which supports the specification of query
subjects. The second one is the word “and”, which indicates that all the keywords
connected by the word are required to have connections with the search results.
The last one is the word “or ”, which supports the specification of optional
keywords. Thus, a user query in SemSearch looks like “subject:keyword1 and/or
keyword2 and/or keyword3 ...”.
With this query syntax, the example of “news about phd students” can be
easily specified as news:phd students, where the term news is the query subject
and the term phd students is a required keyword.
More complex queries in whichSemSearch: A Search Engine for the Semantic Web 7
multiple keywords (except the subject keyword) are involved also can be easily
specified. For example, when querying for projects in which both Enrico and
John participate, the query can be specified as project:Enrico and John. If the
user wants to know all the projects that Enrico or John are involved in, the
query is project:Enrico or John.
In order to satisfy a query, each of the required keywords (please note that
subject is a special required keyword) must be satisfied, which means that the
search results must be semantically related to each of the required keywords. In
the case when there are optional keywords, one of the optional keywords must
be satisfied.
The SemSearch query interface provides a simple, flexible, and powerful ap-
proach for specifying user queries in semantic search. First, it overcomes the
problem of knowledge overhead suffered in formal query fronted search engines
and form-based semantic search engines, as it does not require end users to be
familiar with any particular ontology, semantic data, or any special query lan-
guage. Second, the query interface provides a more flexible way of specifying
queries than the interface presented by form-based search engines. Indeed, it
does not confine users to any pre-defined query subjects and values. Third, in
contrast with current semantic-based keyword search engines which only accept
one keyword as input, this query interface supports the specification of complex
queries in the format of specifying multiple keywords and the expected type of
results. Finally, this query interface is simpler than question answering tools as
the search engine does not need to spend time calculating which of the keywords
are in a user’s query.
In this paper, we focus on the user queries in which there are at least two
keywords involved, in order to better explain the distinctive features of our
search engine. Regarding those queries that only comprise one keyword, the
search engine develops an approach that is similar to the ones used in TAP and
in to find instances that are closely related to the semantic entity matches of
the keyword.