CAJ | 학술논문

不同用户的网络状况、异构环境和地理位置的不同导致其对相同Web服务感受到的QoS差异较大，因此提出了一种为用户预测其未调用过服务的QoS的混合协同过滤方法。首先采用皮尔逊相关系数计算服务之间的相似度，并引入相似度信任因子对相似度进行修正，然后得到与目标服务最相似的top－k服务近邻集，最后采用改进的协同过滤方法预测目标服务的QoS值。采用网络上真实服务的QoS数据集进行实验评估，实验结果表明，该方法可有效提升QoS预测的准确度。
불동용호적망락상황、이구배경화지리위치적불동도치기대상동Web복무감수도적QoS차이교대，인차제출료일충위용호예측기미조용과복무적QoS적혼합협동과려방법。수선채용피이손상관계수계산복무지간적상사도，병인입상사도신임인자대상사도진행수정，연후득도여목표복무최상사적top－k복무근린집，최후채용개진적협동과려방법예측목표복무적QoS치。채용망락상진실복무적QoS수거집진행실험평고，실험결과표명，해방법가유효제승QoS예측적준학도。
The QoS of web service may vary widely for different users due to their different network conditions, heterogeneous environments and locations.In this paper, a hybrid collaborative filtering approach was proposed for QoS prediction.Firstly, Pearson Correlation Coefficient was employed to measure the similarity between two services to find the top-k similar neighbors for target service.Then, the similarity was modified by employing similarity trustworthy factor.Finally, the traditional collabora-tive filtering method was modified to predict the QoS of target service.Real world QoS data were obtained to evaluate the ap-proach.The experimental results demonstrate that the prediction accuracy of QoS could be effectively improved.