Scribd
Upload
4 views2 pages

Human-Provided Services in Expert Networks

This paper discusses the integration of Human-Provided Services (HPS) within service-oriented systems, focusing on the ExpertHITS algorithm for estimating user reputation based on context-sensitive interactions. It highlights the importance of trust and expertise in delegating tasks among peers in the Expert Web, where user ratings influence the effectiveness of service delegation. The implementation and evaluation of the proposed approach demonstrate its scalability and effectiveness in enhancing service interactions.

Uploaded by

praveen kumar
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
4 views2 pages

Human-Provided Services in Expert Networks

This paper discusses the integration of Human-Provided Services (HPS) within service-oriented systems, focusing on the ExpertHITS algorithm for estimating user reputation based on context-sensitive interactions. It highlights the importance of trust and expertise in delegating tasks among peers in the Expert Web, where user ratings influence the effectiveness of service delegation. The implementation and evaluation of the proposed approach demonstrate its scalability and effectiveness in enhancing service interactions.

Uploaded by

praveen kumar
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd

INTRODUCTION

In this paper we utilize Human-Provided Services (HPS) enabling flexible


interactions in serviceoriented systems. We discuss the discovery and interactions
in mixed service oriented systems comprising HPS and software-based services
(SBS). Experts offer their skills and capabilities as HPS that can be requested on
demand. In this work we present the following key contributions: (i) Estimation of
user reputation based on a context-sensitive algorithm. Our approach, called
ExpertHITS, is based on the concept of hubs and authorities in Web-based
environments. (ii) An approach for community reputation (the hub-expertise of
users) influenced by trust relations. Dynamic link weights are based on trust and
user rating influenced by the query context. ExpertHITS is calculated online, thus
fully personalized based on the expert-requester’s preferences (i.e., the demanded
set of skills). (iii) Implementation and evaluation of our approach demonstrating
scalability and effectiveness of our proposed algorithm.

Applying this idea in our scenario, a member of the Expert Web may receive
an RFS and delegate work to some other peer in the network (characterizing hubs
in the network). Receivers of the delegated work, however, expect RFSs fitting
their skills and expertise (i.e., being an authority in the given domain). Careless
delegations of work will overload these peers resulting in degraded processing time
due to missing expertise. Within the ExpertWeb, authorities give feedback using
rating mechanism (e.g., a number on the scale from 1 to 5) to indicate their
satisfaction — whether a particular hub distributes work according to their skills
and interest. Thus, a ‘good hub’ is characterized by a neighborhood of peers that
are satisfied with received RFSs. On the other hand, delegation of work is strongly
influenced by trust, for example, whether the initial receiver of the RFS (hub
within the Expert Web) expects that the peer will process work in a reliable and
timely manner. RFS receivers need to be trusted by influential hubs that are highly
rated in order to be recognized as authoritative peers in the Expert Web. Notice,
hub and authority scores are available for each member in the network. Thus, a
member may act as a hub and authority by processing or delegating tasks.

You might also like