Rap rap is short for reputation aware prediction

Info icon This preview shows pages 15–17. Sign up to view the full content.

. RAP: RAP is short for reputation-aware prediction model, and proposed by Qiu et al. [ 17 ]. This model uses the QoS records to compute the credibility of each user, and then excludes the unreliable users according to the credibility ranking. The missing QoS value is predicted with CF algorithm using the QoS records of reliable users. . LBR2: This method focused on capturing geographical connectivity to identify similar users, and then combine these regularization terms in MF framework [ 15 ]. Table 4 gives the experimental results of all models. We can make the following observations as follows. . The MAE and RMSE values achieved by our proposed model JLMF are consis- tently smaller than baseline models in all settings of training set densities (5% 20%). Let us see a detailed example, the JLMF model achieves the improvement in MAE of 3.27%, 8.53%, 24.06% than LBR2 model, MF model and WSRec model. Besides, the two individual models, ULMF and SLMF, also achieve superior results than the baselines. Such an improvement veri¯es the following points. (1) Our intuition for modeling user and service network location is right. That is, users or services in the same network location are indeed likely to have similar network con¯guration, and further to receive similar QoS values. (2) The location-aware neighbor selection indeed can identify those users (or services) in the same network location. Also, we can infer that the ASs are also useful. QoS Prediction for Web Service Recommendation with Network Location-Aware Neighbor Selection 625 Int. J. Soft. Eng. Knowl. Eng. 2016.26:611-632. Downloaded from by WEIZMANN INSTITUTE OF SCIENCE on 07/01/16. For personal use only.
Image of page 15

Info icon This preview has intentionally blurred sections. Sign up to view the full version.

(3) The way of generating QoS values in our models are e®ective. That is, a QoS value is generated by two parts. One part is learned by the target user's (or service's) QoS records, and the other part is learned by the neighbors' QoS records. (4) The prediction accuracy can be further improved, if a model can fully leverage the user network information and service network information, just as our proposed model JLMF does. . The MAE and RMAE values tend to be smaller as the TD becomes larger. Such a trend means that more user information and service information bene¯t the pre- diction performance. . The improvements achieved by our model JLMF are signi¯cant compared to other baselines, according to the paired t -test ( p < 0 : 001 Þ . In the following experiments, we will study the sensitivity of our proposed model JLMF to the parameters. By observing the experimental results, we ¯nd that the sensitivity of the SLMF model and ULMF model to the parameters is similar to that of the JLMF model. So we report the result of JLMF here. 5.5. The sensitivity to μ The parameter ² controls the proportions of two individual models in the ¯nal predicted QoS value in the JLMF model. We investigate the sensitivity of JLMF to ² in the value range of 0 to 1. The experiments are conducted under the default parameter settings for other parameters, and the training set densities are 10% and 15%. The experimental results are shown in Fig. 3 .
Image of page 16
Image of page 17
This is the end of the preview. Sign up to access the rest of the document.
  • Winter '15
  • MAhmoudali
  • Analysis of algorithms, Computational complexity theory, Service system, Expectation-maximization algorithm, Weizmann Institute of Science

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern