1508 J. Serrano-Guerrero et al Information Sciences 181 (2011)1503-1516 As it was pointed out in [21 this family of transformation functions is bijective. This result guarantees that the transforma- tions between levels of a linguistic hierarchy are carried out without loss of information. To define the computational model we select a level to make the information uniform(for instance, the greatest granularity level)and then we can use the oper ators defined in the 2-tuple FLM 3. Recommender system for the dissemination of information in University Digital Libraries University Digital Libraries allow storing a large amount of digital resources to be accessed by the university communit The researchers may be interested in several topics but they are not expert in every topic that they are interested in. For this reason, a platform is necessary in order to interact with other experts that can support them. The Google waves are useful ool for them since they can access information written by other users, specially useful for new participants, or they can start the interaction with other experts in order to find collaborators or simply to find the answers to their questions. Indeed, the information stored in the waves can be accessed as past experiences in order to know the steps followed in other collabo- ations and to avoid past errors. Therefore the Google waves can be an interesting collaborative tool but it is necessary to have a process which suggests information about people and resources that can participate in the wave because the wave creator/administrator does not know all the resources or people registered into the library that can be useful for the wave. The objective of this section is the description of a recommender system that is able to inform the user about the existence of an interesting wave and to inform the wave administrator about the existence of useful resources that can be included in the wave The system differentiates two kinds of users in the system: administrators and users. Both administrators and users are researchers who use the University Digital Library but can play different roles. The administrators are experts interested in finding collaborators for research purposes and hence they are in charge of defining the characteristics of a wave. Moreover they are in charge of inserting the resources because they know best the characteristics of each resource On the other hand, users receive requests for collaboration and decide if they want to participate in the wave. therefore the activity of a researcher consists only in defining his profile and participating in the chosen waves. The users of the wave an propose to the administrator the inclusion of new resources that they have and know would be useful for the wave users but the administrator is who ultimately decides if the resource is inserted or not. Our system presents three main recommendations depending on the action that is carried out: Insertion of a user. When a user is logged into the system he receives information about the waves which he might be Insertion of a digital resource. When a digital resource is inserted into the system, the administrators of the system receive a recommendation about which waves can include the resource Insertion of a wave. when a new wave is created the system searches people and resources that may be useful for the The insertion of each element (users, resources and waves) consists in filling out a form which contains the information that characterizes each element That information is described in the following subsection. Each form presents a set of label ts(S1, S2,... that allows the assessment of the different concepts managed by system. These label sets are chosen from the label sets that compose a Linguistic Hierarchy. It is necessary to remark that the number of levels of the Linguistic Hierarchy restricts the number of different label sets, therefore two label sets Si and Sy can represent the label set of the Linguistic Hier archy however the interpretation of both(SSy)can be different. In the proposed system it is necessary to distinguish four concepts that can be assessed Importance degree (Sn)of a discipline with respect to the scope of the resource the user preferences or the scope of the Importance degree(S,)of a wave according to user opinion. Importance degree(S3)of a resource for a wave. Importance degree(S4)of a resource according to the opinion of the person who required the insertion of that resource llowing the linguistic hierarchy shown in Fig. 2, in our system we use level 2 (5 labels )to assign importance degree (S1=S)and level 3(9 labels)to assign importance degrees(S2=S3=S4=S). Using this lh the linguistic terms in each levelAs it was pointed out in [21] this family of transformation functions is bijective. This result guarantees that the transforma￾tions between levels of a linguistic hierarchy are carried out without loss of information. To define the computational model, we select a level to make the information uniform (for instance, the greatest granularity level) and then we can use the oper￾ators defined in the 2-tuple FLM. 3. Recommender system for the dissemination of information in University Digital Libraries University Digital Libraries allow storing a large amount of digital resources to be accessed by the university community. The researchers may be interested in several topics but they are not expert in every topic that they are interested in. For this reason, a platform is necessary in order to interact with other experts that can support them. The Google waves are useful tool for them since they can access information written by other users, specially useful for new participants, or they can start the interaction with other experts in order to find collaborators or simply to find the answers to their questions. Indeed, the information stored in the waves can be accessed as past experiences in order to know the steps followed in other collabo￾rations and to avoid past errors. Therefore the Google waves can be an interesting collaborative tool but it is necessary to have a process which suggests information about people and resources that can participate in the wave because the wave creator/administrator does not know all the resources or people registered into the library that can be useful for the wave. The objective of this section is the description of a recommender system that is able to inform the user about the existence of an interesting wave and to inform the wave administrator about the existence of useful resources that can be included in the wave. 3.1. Roles The system differentiates two kinds of users in the system: administrators and users. Both administrators and users are researchers who use the University Digital Library but can play different roles. The administrators are experts interested in finding collaborators for research purposes and hence they are in charge of defining the characteristics of a wave. Moreover they are in charge of inserting the resources because they know best the characteristics of each resource. On the other hand, users receive requests for collaboration and decide if they want to participate in the wave, therefore the activity of a researcher consists only in defining his profile and participating in the chosen waves. The users of the wave can propose to the administrator the inclusion of new resources that they have and know would be useful for the wave users, but the administrator is who ultimately decides if the resource is inserted or not. 3.2. Recommendations Our system presents three main recommendations depending on the action that is carried out: Insertion of a user. When a user is logged into the system he receives information about the waves which he might be interested in. Insertion of a digital resource. When a digital resource is inserted into the system, the administrators of the system receive a recommendation about which waves can include the resource. Insertion of a wave. When a new wave is created, the system searches people and resources that may be useful for the waves purposes. The insertion of each element (users, resources and waves) consists in filling out a form which contains the information that characterizes each element. That information is described in the following subsection. Each form presents a set of label sets (S1,S2,...) that allows the assessment of the different concepts managed by system. These label sets are chosen from the label sets that compose a Linguistic Hierarchy. It is necessary to remark that the number of levels of the Linguistic Hierarchy restricts the number of different label sets, therefore two label sets Si and Sj can represent the label set of the Linguistic Hier￾archy however the interpretation of both (Si,Sj) can be different. In the proposed system it is necessary to distinguish four concepts that can be assessed: Importance degree (S1) of a discipline with respect to the scope of the resource, the user preferences or the scope of the wave. Importance degree (S2) of a wave according to user opinion. Importance degree (S3) of a resource for a wave. Importance degree (S4) of a resource according to the opinion of the person who required the insertion of that resource. Following the linguistic hierarchy shown in Fig. 2, in our system we use level 2 (5 labels) to assign importance degree (S1 = S5 ) and level 3 (9 labels) to assign importance degrees (S2 = S3 = S4 = S9 ). Using this LH the linguistic terms in each level are: 1508 J. Serrano-Guerrero et al. / Information Sciences 181 (2011) 1503–1516