南京师大学报(自然科学版)
南京師大學報(自然科學版)
남경사대학보(자연과학판)
JOURNAL OF NANJING NORMAL UNIVERSITY (NATURAL SCIENCE EDITION)
2014年
1期
22-29,46
,共9页
QoS%Web服务%信任度%声誉%推荐信任%偏好
QoS%Web服務%信任度%聲譽%推薦信任%偏好
QoS%Web복무%신임도%성예%추천신임%편호
QoS%web service%reliability%reputation%recommended trust%preferences
Web服务的可信性往往对面向服务的应用起决定性作用。由于服务的信任信息来源很多,整合多种来源的服务信任信息,准确地评估服务的可信性是一个挑战。针对这一挑战,提出了一种基于QoS和声誉的服务信任度评估方法。与其他服务信任度评估模型相比,本模型融合了多个信任信息来源,在直接信任度中综合了QoS和用户对服务的历史信任,并加入时间衰减因子,逐渐削弱历史信任对直接信任度的影响,提高直接信任度的可靠性;在推荐信任度中引入推荐者声誉概念,建立用户之间的信任评估机制,从而将推荐信任分为熟人推荐信任和陌生人推荐信任两类,使推荐信任的评估更加完整、准确。此外,本方法在时间和空间复杂度方面,也存在诸多优势。实例分析表明,本文提出的评估方法可以用来解决Web服务选择的可信问题,并且能够提高服务信任度评估的准确性。
Web服務的可信性往往對麵嚮服務的應用起決定性作用。由于服務的信任信息來源很多,整閤多種來源的服務信任信息,準確地評估服務的可信性是一箇挑戰。針對這一挑戰,提齣瞭一種基于QoS和聲譽的服務信任度評估方法。與其他服務信任度評估模型相比,本模型融閤瞭多箇信任信息來源,在直接信任度中綜閤瞭QoS和用戶對服務的歷史信任,併加入時間衰減因子,逐漸削弱歷史信任對直接信任度的影響,提高直接信任度的可靠性;在推薦信任度中引入推薦者聲譽概唸,建立用戶之間的信任評估機製,從而將推薦信任分為熟人推薦信任和陌生人推薦信任兩類,使推薦信任的評估更加完整、準確。此外,本方法在時間和空間複雜度方麵,也存在諸多優勢。實例分析錶明,本文提齣的評估方法可以用來解決Web服務選擇的可信問題,併且能夠提高服務信任度評估的準確性。
Web복무적가신성왕왕대면향복무적응용기결정성작용。유우복무적신임신식래원흔다,정합다충래원적복무신임신식,준학지평고복무적가신성시일개도전。침대저일도전,제출료일충기우QoS화성예적복무신임도평고방법。여기타복무신임도평고모형상비,본모형융합료다개신임신식래원,재직접신임도중종합료QoS화용호대복무적역사신임,병가입시간쇠감인자,축점삭약역사신임대직접신임도적영향,제고직접신임도적가고성;재추천신임도중인입추천자성예개념,건립용호지간적신임평고궤제,종이장추천신임분위숙인추천신임화맥생인추천신임량류,사추천신임적평고경가완정、준학。차외,본방법재시간화공간복잡도방면,야존재제다우세。실례분석표명,본문제출적평고방법가이용래해결Web복무선택적가신문제,병차능구제고복무신임도평고적준학성。
The credibility of web service often plays an important role for service-oriented applications. Integrating and assessing different sources of trust is a challenge. In this paper,we present a method of web service reliability assessment based on QoS and reputation. We not only incorporate many sources of trust,but also assess the web service' trust. Firstly, we integrate direct trust with multi-QoS of web services and the historical trust of the users and services. At the same time, the reliability of direct trust is improved through the addition of the attenuation factor to weaken the influence of the historical trust to the direct trust. Secondly, we present the concept of reputation and establish trust assess mechanism between users. Then recommended trust is divided into acquaintances' indirect recommendation trust and strangers' direct recommendation trust,so as to make the recommendation trust value more comprehensive and credible. The method is proved to be better than other service's trust model in time and space complexity. The result of example indicates that the proposed method can be used to solve the trusted web service selection,and guarantee the accuracy of the trust.