Recommender Service. (GMD) This Task will produce the internal architecture, the detailed specification and the implementation of the system component which supports the recommender service. This service is responsible for providing recommendations to satisfy the information needs of a user by exploiting both user and community profiles. These recommendations could be delivered on user demand or asynchronously, i.e., periodically. In particular the following functionality will be supported: community building, rating, query expansion, recommendation. The core of the RS will be the hybrid space concept. In hybrid space, users are partitioned in neighbourhoods. Each neighbourhood will identify a community of interest in which users will share the same tastes and information needs. The hybrid space concept will be developed by integrating two models: the social space model and the content space model. The former model will identify a user community on the basis of the similarity between the rating pattern of a user with those of other users. The latter will identify a user community on the basis of the similarity between the content-based profile of a user and that of other users. The hybrid space concept will permit the co-operation between these two models. The activities under this Task have a strong research orientation. Task T3.6
Recommender Service. CGRA evaluators used the service and completed the related Dicode Evaluation Framework Questionnaires. Figures 3.2.5.1, 3.2.5.2, 3.2.5.3 summarize the evaluators’ responses relative to the quality, acceptability, accessibility and usability of the Recommender service. In these figures, the median and mode values of the evaluators’ responses are shown, where those were given in a 1-5 scale. Regarding the overall quality of the service the evaluators’ responses are neutral concerning whether the service addressed the data intensive decision making issues, if the objectives of the service are met, if the service is novel to their knowledge, and if overall they are satisfied with the service. Comments such as ‘seems to help but depends on the application and the framework.’ (GR16), ‘I don't know the objectives in detail but I think they are met’ [sic] (GR32) and ‘Its perfomance was very good, its responce time too, the only thing missing was a better user interface.’ [sic] (GR12), are helpful in explaining the just above the average median and mode score values in the overall quality of the service. The evaluators agree that they are satisfied with the performance of the service (median and mode values of 4). Regarding the acceptability of the service, as can be seen in Figure 3.2.5.2, the evaluators are neutral in that the service is helpful to their work, it has all the functions they expected, the interface of the service was pleasant, they will use this service in their work and that they will recommend this service to their peers/community. The acceptability of the service perhaps can be explained by the following comment: ‘The function of the service is very interesting, I cannot see how it could be extended.’ (GR31). In Figure 3.2.5.3 we can observe the accessibility and other related issues median and mode score values given for the Recommender service. The evaluators agree with the access given, the instruction given by the service, they are satisfied with the time spend to learn to use the provided solution and they are satisfied with the ease of use of the solution to complete their tasks. Evaluators agreed that the service was easy to use and fast, although a noticeable comment refers to its output: ‘Wasn't easy due to the complex outputs.’ (UK1). Reviewers also agree in the lack of interface. The system’s messages provided were thought to be insufficient: ‘I wanted more system messages/help in order to understand the service's use and errors.’ (GR...
Recommender Service. In order to use the service a REST-Client such as RESTClient must be installed (xxxxx://xxxxxx.xxx/chao/RESTClient). Example Request: GET xxxx://xxxxxxxx.xxxx.xxxxxxxxxx.xx:9091/REST/Recommender/Input/Sample Example Output: Please complete the following Questionnaires: Questionnaire 1: Dicode Evaluation-First Trial-Standalone Service-Recommender Service, xxxx://xxxxxxxxxxx.xxx?s=LKLJGI_fa8ccd89 Questionnaire 2: Dicode Evaluation-First Trial-Standalone Service-The Recommender Service -Ease of Use Assessment User Questionnaire xxxx://xxxxxxxxxxx.xxx?s=LKLJGH_8d8bfd1f Questionnaire 3: Dicode Evaluation-First Trial-Standalone Service-The Recommender Service - Acceptability Assessment User Questionnaire xxxx://xxxxxxxxxxx.xxx?s=LKLJGF_6a33d018 Additional instructions in: xxxxx://xxxx.xxxxxx-xxxxxxx.xx/display/DIC/Recommender+Service
