J. Serrano-Guerrero et aL Information Sciences 181(2011)1503-1516 1511 Apart from the vector v the wave definition includes a name, a description, i.e., a short explanation of the purpose of the wave, a list of descriptive words(key-words )about the topic of the wave and a list of administrators, i.e. people in charge of creating and supervising the activity within the wave The wave administrator has to assign a linguistic threshold value for resources(a)which indicates the minimum value lat can be required by an administrator for informing him about the existence of a new interesting resource for his admin- strated waves. Indeed the user has to fix a linguistic threshold value for users(o) which determines the minimum relation- Finally the wave definition is completed by a list of users who interact with the wave and a list of resources available in the wave. Each user describes the degree of importance bx E S3 of the wave with respect to his interests as was discussed in Section 3.3. 2, and furthermore, the user who suggested the resource or the wave administrator describes the importance degree bx E S] of that resource for the wave as was discussed in Section 3.3.1 3.4. Steps of the system he proposed approach, contrary to other works 40, 41, does not suggest collaborators or resources for a particular user but recommends the activity of users and resources within specialized forums through the collaborative tool called Google Wave. a detailed explanation of the recommendation process is described in the following subsections 3.4.1. Insertion of a wave When several researchers decide that it is necessary to open a new wave in order to find people and resources for a new initiative, they communicate their desire to the library staff and then the staff members decide if there are enough research- ers for creating the wave and if the proposal is relevant. If the request is accepted, one or various researchers are selected a administrators and are in charge of inserting the first users and resources into the wave by hand because they know of ticipants interested in the wave and the resources that they have. when this process is finished the system starts with automatic recommendations. There are two possible elements to be suggested for each wave i, a set of users Ui and a set of resources R a general view of the system can be seen in Fig 4 within a digital library with many other tools. The flow of the recommendations for the wave is depicted in Fig. 5. Each user has to be registered into the system by filling out the registration forms whereas the resources are registered by the administrators of the wave filling out other form as will be discussed below. The definition of the waves is performed by one or several administrators and completed by the recommender system which associates the most relevant resources and users with the wave. The first members and resources of the wave are inserted by hand, but the rest of the process is computed automatically The internal representation of the wave i is a vector Vw which stores 2-tuple linguistic values for each category defined in Science Direct apart from the other data commented in Section 3.3.3. The system matches the wave vector wi and the vec tors VUk for each user k within the system and if the result exceeds the threshold o fixed by the wave administrator/s (oNV VUk)>o) then the system informs the user k via email about the existence of the wave i Next, users have to decide Administrator Resource vectors wave vector Matc User Fig. 5. Flow of the recommender system for the wave definition.Apart from the vector VW the wave definition includes a name, a description, i.e., a short explanation of the purpose of the wave, a list of descriptive words (key-words) about the topic of the wave and a list of administrators, i.e., people in charge of creating and supervising the activity within the wave. The wave administrator has to assign a linguistic threshold value for resources (k) which indicates the minimum value that can be required by an administrator for informing him about the existence of a new interesting resource for his admin￾istrated waves. Indeed, the user has to fix a linguistic threshold value for users (x) which determines the minimum relation￾ship between the wave and an existing user in the system in order to consider that both share interests. Finally the wave definition is completed by a list of users who interact with the wave and a list of resources available in the wave. Each user describes the degree of importance bx 2 S3 of the wave with respect to his interests as was discussed in Section 3.3.2, and furthermore, the user who suggested the resource or the wave administrator describes the importance degree bx 2 S2 of that resource for the wave as was discussed in Section 3.3.1. 3.4. Steps of the system The proposed approach, contrary to other works [40,41], does not suggest collaborators or resources for a particular user but recommends the activity of users and resources within specialized forums through the collaborative tool called Google Wave. A detailed explanation of the recommendation process is described in the following subsections. 3.4.1. Insertion of a wave When several researchers decide that it is necessary to open a new wave in order to find people and resources for a new initiative, they communicate their desire to the library staff and then the staff members decide if there are enough research￾ers for creating the wave and if the proposal is relevant. If the request is accepted, one or various researchers are selected as administrators and are in charge of inserting the first users and resources into the wave by hand because they know of par￾ticipants interested in the wave and the resources that they have. When this process is finished the system starts with the automatic recommendations. There are two possible elements to be suggested for each wave i, a set of users Ui and a set of resources Ri. A general view of the system can be seen in Fig. 4 within a digital library with many other tools. The flow of the recommendations for the wave is depicted in Fig. 5. Each user has to be registered into the system by filling out the registration forms whereas the resources are registered by the administrators of the wave filling out other form as will be discussed below. The definition of the waves is performed by one or several administrators and completed by the recommender system which associates the most relevant resources and users with the wave. The first members and resources of the wave are inserted by hand, but the rest of the process is computed automatically. The internal representation of the wave i is a vector VWi which stores 2-tuple linguistic values for each category defined in ScienceDirect apart from the other data commented in Section 3.3.3. The system matches the wave vector VWi and the vec￾tors VUk for each user k within the system and if the result exceeds the threshold x fixed by the wave administrator/s (rl(VWi,VUk) > x) then the system informs the user k via email about the existence of the wave i. Next, users have to decide User insertion Resources insertion Administrator User User vectors Resource vectors wave vector Matching resources users Wave definition Resources Management Recommendations Users Management Waves Management wave vector Fig. 5. Flow of the recommender system for the wave definition. J. Serrano-Guerrero et al. / Information Sciences 181 (2011) 1503–1516 1511