Services are software components with standard interfaces that enable rapid and flexible development of distributed systems over the Web. The literature proved that service matchmaking is the most effective approach for finding services that fulfil user needs. Available matchmaking tools require a repository composed of semi-structured descriptions that specify service properties according to a shared model. The diffusion of services that provide public functionalities on the Web, called Web APIs, is leading to a new scenario in which increasing Web information about services is available on dispersed sources, such as official documentations, wikis, forums, blogs and social networks. In order to exploit Web information to support service matchmaking, several issues must be addressed. Web sources provides service information according to heterogeneous data formats, models and vocabularies. Web information is dynamic, then can change over time. Available descriptions can be invalid in terms of accuracy, currency and trustworthiness. Sources provide partial or contradictory information about same services. Finally, some relevant properties, such as service provider popularity or quality of service documentation, are not explicitly reported on the Web, then may be wrongly interpreted by users through personal subjective evaluations. This thesis provides an overall approach for enabling effective and efficient service match- making on Web information by addressing the issues above. The approach constructs semantic descriptions by extracting (i) explicit property values from heterogeneous non- semantic sources; (ii) subjective property values through social media. Then, quality assessments on Web information are exploited for fusing valid descriptions extracted by several sources and performing effective matchmaking. The overall approach is implemented through a lightweight distributed architecture which focuses on scalability issues by managing big amount of Web information about services.
(2013). Service matchmaking: exploiting the Web. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2013).
Service matchmaking: exploiting the Web
PANZIERA, LUCA
2013
Abstract
Services are software components with standard interfaces that enable rapid and flexible development of distributed systems over the Web. The literature proved that service matchmaking is the most effective approach for finding services that fulfil user needs. Available matchmaking tools require a repository composed of semi-structured descriptions that specify service properties according to a shared model. The diffusion of services that provide public functionalities on the Web, called Web APIs, is leading to a new scenario in which increasing Web information about services is available on dispersed sources, such as official documentations, wikis, forums, blogs and social networks. In order to exploit Web information to support service matchmaking, several issues must be addressed. Web sources provides service information according to heterogeneous data formats, models and vocabularies. Web information is dynamic, then can change over time. Available descriptions can be invalid in terms of accuracy, currency and trustworthiness. Sources provide partial or contradictory information about same services. Finally, some relevant properties, such as service provider popularity or quality of service documentation, are not explicitly reported on the Web, then may be wrongly interpreted by users through personal subjective evaluations. This thesis provides an overall approach for enabling effective and efficient service match- making on Web information by addressing the issues above. The approach constructs semantic descriptions by extracting (i) explicit property values from heterogeneous non- semantic sources; (ii) subjective property values through social media. Then, quality assessments on Web information are exploited for fusing valid descriptions extracted by several sources and performing effective matchmaking. The overall approach is implemented through a lightweight distributed architecture which focuses on scalability issues by managing big amount of Web information about services.File | Dimensione | Formato | |
---|---|---|---|
phdthesis_panziera.pdf
accesso aperto
Tipologia di allegato:
Doctoral thesis
Dimensione
1.7 MB
Formato
Adobe PDF
|
1.7 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.