RU2461150C1 - Method and apparatus for selecting and indicating service - Google Patents

Abstract

FIELD: information technology.

SUBSTANCE: method involves providing a data structure representing user terminals and relationships between the terminals, obtained from one or more communication parameters associated with communication between user terminals. Also, information on service use for each of a plurality of services is provided for at least some of the user terminals. The method further involves selecting, for at least one user terminal and from a plurality of services not used by at least one user terminal, a service which is not used by at least one user terminal, wherein the service selection is based on information of service use provided for user terminals, having predetermined connection with at least one user terminal, as a defined relationship between terminals, and sending a message indicating the selected service to at least one user terminal.

EFFECT: efficient indication of service accessibility in a user terminal in a communication network.

19 cl, 11 dwg

Description

FIELD OF THE INVENTION

The disclosure generally relates to communication networks. Specifically, the disclosure relates to the technical field of providing guidance regarding the availability (including availability) of services to a user terminal in a communication network.

State of the art

In modern communication networks, there is an abundance of various communication services. Available communications services include traditional telecommunications services such as voice services, data services, short message service (SMS), multimedia messaging service (MMS). In addition, instant messaging (IM) and group-based communication services (typically based on the Internet or IP protocol), possibly including more than two communication partners, are becoming more and more popular. IP-based group services include, for example, blogging, virtual forums, etc.

As a result of the large number of different communication services currently in use, the communication environment as a whole has become quite heterogeneous. The heterogeneous nature of the communication environment is expressed by the fact that many of the communication services and, in particular, group-based communication services are not standardized. Also, different user terminals (such as different mobile phones) often support different sets of communication services, which means that in connection with a particular communication service, two user terminals can sometimes be incompatible with each other. Thus, the heterogeneous nature of the communication medium to a certain extent impedes the possibility of communication between user terminals.

In other circumstances, the ability to communicate between user terminals is simply limited, as a particular user may not be aware of a particular communication service supported by his or her user terminal in his current configuration or after a quick software update or subscription to the service. In other words, a user may have more communication options using a user terminal than those that he or she actually knows about. For example, if the user terminal supports traditional communication services via a cellular communication network (for example, in accordance with the specifications of the universal mobile telecommunication system, or UMTS) and, in addition, group-based communication services (which can also be provided by a wireless LAN or WLAN ), the user can only know about the functionality of the UMTS of his or her terminal. As a result, the user will not be given the opportunity of any communication in areas having only a WLAN service area, despite the fact that the terminal can support access to the network via WLAN. The lack of communication capabilities that follows is particularly annoying if many of the potential user communication partners regularly use such group-based communication services. On the other hand, in many situations it may be useless to notify the user of supported communication services that he or she will not use, since there are not or are only a few possible communication partners.

What was mentioned above in connection with communication services also applies to certain application services. That is, the user in many cases can benefit from applications that are not currently used, which, in principle, are supported by his or her terminal.

SUMMARY OF THE INVENTION

What is needed is a method that more efficiently provides indications of service availability to a user terminal in a communication network.

In accordance with a first aspect, a method is provided for providing an indication of service availability to a user terminal. The method comprises providing a data structure representing user terminals and relationships between terminals obtained from one or more communication parameters related to communications between user terminals, providing at least some of the user terminals with information about using the service for each of the plurality of services, selection, for at least one user terminal and from among a plurality of services, a service not used by at least one user minala, moreover, the choice of service is based on information on the use of the service provided for user terminals having a predefined association with at least one user terminal, as determined by the relationship between the terminals, and sending a message indicating the selected service to at least one user terminal. As is understood in the present application, the term “plurality” refers to a number greater than one.

The data structure can be implemented in various forms, such as as an abstract data model or as a set of data sets. The data structure may be configured to be represented or may be represented as a network graph with nodes and edges connecting every two nodes. In a network graph, nodes correspond to user terminals, and edges can be obtained from relationships between nodes. For example, an edge can be used to connect two nodes that represent user terminals between which a communication link has been established in the past. When a network graph is used, the preferred association between user terminals (i.e. nodes of the network graph) underlying the service selection can be determined in terms of features related to the topology of the network graph.

The service selection procedure can be performed in the context of indicating to the user terminal an available service that is not currently used and / or previously by the user terminal. The selection may take into account the technical functionality of a particular user terminal to ensure that only the services actually supported by the user terminal are recommended (including cases in which only software updates or service descriptions are required). A predefined association between a user terminal that forms the target of the recommendation and other user terminals that already use the recommended service can be expressed as one or more criteria. Criteria can directly or indirectly (for example, using the topology of the network graph) relate to the relationship between the terminals, as obtained from one or more communication parameters between the terminals.

In one example, the predefined association underlying the service selection procedure contains a criterion expressed as a measure of distance from the point of view of the network graph or applicable to the network graph. The distance measure can be a predetermined number of “transitions” interconnecting specific nodes in the network graph. In one embodiment, the choice of service, for example, may be based on information about the use of the service provided for user terminals that are direct neighboring terminals of the user terminal to which the indication of service availability should be transmitted. In another implementation, the choice of service can be based on information about the use of the service provided for user terminals, which can be achieved using a predefined maximum number of transitions (such as, for example, one, two or three transitions) from the user terminal, which forms the goal of indicating availability services.

In a network graph, each edge can be associated with metric information that quantifies the relationship between two nodes connected by a particular edge. Information about the metrics can be at least partially obtained from one or more communication parameters (or, which in many cases will be equivalent, from relations between terminals). Alternatively, or additionally, metric information can be obtained from other parameters, such as the topology of the network graph. Parameters obtained from the topology of the network graph include, for example, the presence of groupings or the nature of transitive relations from the perspective of a particular node. In cases in which metric information is provided, the predefined association underlying the service selection procedure may contain a criterion expressed in terms of metric information or applicable to metric information.

The edges in the network graph can be directed (i.e., they can indicate a predefined direction, shown, for example, using an arrow). Directional edges can be used in various contexts, for example, to express the level of trust between users of different user terminals related between the terminals. Some common characteristics of trust include that it is transitive and asymmetric (i.e., trust is usually not bidirectional). In the case of directed ribs, the predefined association underlying the service selection procedure may contain a criterion expressed in terms of the direction of the rib, or applicable to the direction of the rib.

At least some of the nodes in the network graph (or user terminals in the data structure) can be ranked, and then the predefined association underlying the service selection procedure may contain a criterion expressed in terms of ranking. In one embodiment, the nodes are ranked based on at least one of the following: their membership in the grouping and the presence of transitive relationships from the point of view of the network graph.

For a particular edge in a network graph, individual metric information elements may be provided for each of a plurality of communication services. In such a scenario, an assessment of whether or not to provide an indication of service availability in connection with a particular communication service is performed based on metric information provided for that specific communication service. In other words, the evaluation of different communication services in connection with one single rib will include different elements of information about the metrics.

In accordance with the first embodiment, information about the use of the service relates to communication services. Thus, the selected service will be a communication service. Possible communication services include one or more of voice services, data services, multimedia services, text services, instant messaging services and group based communication services. In another embodiment, service usage information relates to application services, and thus, the selected service will be an application service. Possible application services include, but are not limited to, mobile application services, such as mobile gaming applications, potentially causing the participation and interaction of multiple mobile user terminals, and a location determiner application (i.e., applications directing the user managing the mobile terminal, to a specific place). Application services can be services that are provided by a service providing network to which user terminals can connect using a communication service or a communication network.

In one scenario, individual services are ranked according to service usage information. This means, for example, that frequently used services rank higher than services that only occasionally use. In other words, the ranking of services may depend on the use of each individual service by a separate user terminal, terminals included in the data structure, or large collections of terminals. In this case, the choice of service may also be based on the ranking of services.

One or more communication parameters underlying the relationship between the terminals may include one or more of the following: communication partner (s), communication duration, number of communications, communication direction and type of communication service used. Of course, the relationship between the terminals can also be obtained, taking into account other parameters that are not directly related to the communication between the two user terminals (such as routing messages from the first user terminal through the second user terminal to the third user terminal). There are various options for obtaining communication parameters. For example, communication parameters can be obtained from at least one of the following: call detail records, event detail records, investigation of a communication network interface, and user terminal agents.

In accordance with a further aspect, a computer program product is provided that comprises portions of program code for performing steps of one or more methods and aspects of the methods described herein when the computer program product is executed in one or more computing devices. The computer program product may be stored on a computer-readable medium, such as read-only or write-back memory in a computing device or associated with a computing device, CD, DVD, etc. In addition, or alternatively, a computer program product may be provided for downloading to a computing device, for example, using a data network such as the Internet, or a dedicated communication line such as a telephone line or a wireless line.

In accordance with yet another aspect, an apparatus is provided for providing an indication of service availability to a user terminal. The device comprises a memory providing a data structure representing user terminals and relationships between terminals obtained from one or more communication parameters related to communications between user terminals, and further providing at least for some of the user terminals, service usage information for each of the plurality of services, a processor adapted to select for at least one user terminal and from a variety of services of a service not used at least one user terminal, the service selection being based on service usage information provided for user terminals having a predetermined association with at least one user terminal, as determined by the relationship between the terminals, and an output interface for sending a message, indicating the selected service to the user terminal. The memory may be located in one place (for example, in the form of a single database). In an alternative implementation, the memory is distributed between two or more places (for example, in the form of two or more specialized databases).

In one configuration, the data structure is adapted to be represented (for example, on a computer monitor or any other display device) as a network graph with nodes and edges, in which nodes are representative of user terminals and in which edges connect every two nodes, and they are obtained from the relations between nodes. In a network graph scenario, the predefined association underlying the service selection may contain a criterion expressed as a measure of distance, from the point of view of the network graph, or applicable to the network graph.

The device and its components can be configured to perform one or more of the steps of the methods described in this application.

Brief Description of the Drawings

Further details and aspects will be discussed with reference to illustrative embodiments illustrated in the drawings, in which

figure 1 illustrates an embodiment of a device for providing indications of service availability at user terminals;

2 illustrates an embodiment of a data structure,

figure 3 illustrates a network graph containing nodes and edges obtained from the data structure of figure 2;

4 shows a flowchart schematically illustrating an embodiment of a method;

5 schematically illustrates a network graph containing multiple groupings;

6 schematically illustrates the concept of transitive relationships with respect to an illustrative network graph;

7 schematically illustrates an additional network graph containing several groups;

Fig.8 illustrates a network graph containing directed edges,

Fig.9 illustrates the definition of grouping in a network graph, including directed edges;

10 is a diagram illustrating a comparison between the edge counting mechanism and the group counting mechanism with an average cluster size of 5, and

11 is a diagram illustrating a comparison between the edge counting mechanism and the group counting mechanism with an average cluster size of 20.

DETAILED DESCRIPTION OF THE INVENTION

In the following description, for purposes of explanation and not limitation, specific details are set forth in order to provide a thorough understanding of the methods presented in this application. One skilled in the art will recognize that these methods may be practiced in other embodiments that differ from these specific aspects. For example, one skilled in the art will understand that these methods can be practiced not only in connection with a mobile communication environment, but also in connection with wired communication systems.

One of ordinary skill in the art will further understand that the functions explained in this application below, at least in part, can be implemented using separate hardware circuits, software that operates in conjunction with a programmed microprocessor, programmable gate array (FPGA), integrated circuit Application Orientation (ASIC) and / or Digital Signal Processor (DSP).

An exemplary embodiment of a method for providing indications of service availability will now be described with reference to FIG. FIG. 1 illustrates an illustrative embodiment of an apparatus 100 comprising various subsystems working together in a process for generating and providing indications of service availability for a plurality of 102 user terminals 98.

A plurality of terminals 102 is typically a mixture of mobile and fixed terminals 98 configured to communicate with each other using multiple communication networks and network access methods. A collection of 102 terminals, for example, may include mobile phones, personal digital assistants, data cards or network cards, personal computers, laptops, etc. Communication networks enabling communication between user terminals may include the Internet, mobile or landline telephone networks, etc. Possible network access methods used by user terminals may include wireless network access methods (e.g., code division multiple access or WCDMA methods, and long-term development project or LTE methods, as well as methods for accessing wired networks ( for example, digital subscriber line or DSL methods).

In the exemplary embodiment illustrated in FIG. 1, the complete device 100 comprises three specialized subsystems, namely, a data subsystem 104, a data analysis subsystem 106, and a messaging subsystem 108. The data subsystem 104 comprises components adapted to collect data related to a plurality of user terminals 102. Such data includes communication parameters related to communications between user terminals 98, as well as service usage parameters related to the use of various services (such as communication and application services) by user terminals 98. In the particular case of communication services, communication parameters and usage parameters services, at least in part, may be identical. As one example, the duration of a particular communication session is a communication parameter, which can also be considered as a parameter of the use of the service in connection with the specific communication service underlying the communication session.

The data collected by the data subsystem 104 provide to the data analysis subsystem 106. The data analysis subsystem 106 is configured to evaluate the data provided by the data subsystem 104 in connection with individual user terminals 98 (or, which in many cases will be equivalent in connection with individual subscribers or subscriptions to services). The main purpose of the data analysis subsystem 106 is to identify individual communication services that can increase the connectivity of individual user terminals 98, as well as to identify individual application services that can increase the usability of individual user terminals 98.

As shown in FIG. 1, a data analysis subsystem 106 is connected to a messaging subsystem 108. The messaging subsystem 108 is configured to receive information related to individual user terminals and individual services from the data analysis subsystem 106 and the generation and addressing of messages indicating available services to the individual user terminals 98 of the collection of 102 terminals.

Hereinafter, the components and processing operations of the individual subsystems 104, 106, and 108 will be described in more detail in order to provide a better understanding of the illustrative embodiment illustrated in FIG. In this regard, reference will be made to the schematic data structure 200 of FIG. 2, the illustrative network graph 300 of FIG. 3, and also the flowchart 400 of the flowchart of FIG. 4.

In a first operational step, the data subsystem 104 collects information about a plurality of terminals 102 from communication networks and service networks, such as the IP multimedia subsystem (IMS) used by a plurality of 102 terminals. Such information usually includes various communication parameters, as well as various parameters for the use of services. Communication parameters can be provided for each of a variety of communication services, such as voice services (including traditional telephone services), data services, multimedia services (including MMS), text services (including SMS), IM services and group-based communication services (such like blogging and virtual forums). Communication parameters collected for each of these communication services may include one or more of the following: communication duration, communication partners (for example, in the form of permanent network addresses of participating user terminals 98), the number of communications (for example, the number of SMS messages exchanged two user terminals 98), the direction of communication (for example, unidirectional or bidirectional), and the type of communication service used (for example, telephone service or MMS).

In the case of communication services, service usage parameters may be obtained from one or more communication parameters and optionally from additional information specific to the service, or may contain one or more communication parameters and optionally additional information specific to the service. In the case of application services, the parameters for using the service may contain at least one of the following: the type of application service used (for example, IMS service), the number of uses of the service (for example, the total number of times a specific IMS service has been used), and the total duration of use of the service (for example in minutes or hours).

Information on a plurality of terminals 102 may be collected in various ways. The data subsystem 104 includes, for example, terminal agents 110 located in separate user terminals 98 of the terminal assembly 102 and automatically collects the necessary information and transfers the collected information to the central data acquisition module 112 of the data subsystem 104. Terminal agents 110, for example, can be implemented as specialized software modules running in at least one of user terminals 98, constituting a plurality of terminals 102.

Data collection component 112 can also collect information about a plurality of terminals 102 using call detail records (CDRs) and event detail records (EDRs) that can be obtained from network operator databases or via terminal agents 110. CDR and EDR data can be viewed as lower level data describing user behavior (e.g., duration of a particular communication session) and user relationships (e.g. network addresses, including phone numbers of participating communication partners).

An additional opportunity to obtain information about the totality of 102 terminals is the study of network interfaces using a specialized monitoring component 114, which is connected to the data collection component 112. In one embodiment, the monitoring component 114 is located to monitor traffic in a predefined communication network interface. In an example UMTS network, the monitoring component 114 may be located in a Uu interface, a CS Iu interface, or a PS Iu interface. In another embodiment, the monitoring component 114 is located in the service network interface. In an example IMS network, the monitoring component 114 may be located at the interface between the Call Session Management Function (CSCF) node and the IMS Application Server (AS).

The main goal of the data collection module 112 is to collect the necessary information about the totality of 102 terminals, as well as enrich the collected information with additional data, when necessary. Since the information collected by the data collection module 112 will usually be very heterogeneous, an additional important task of the data collection component 112 is to normalize the collected information. The normalization operation performed by the data collection component 112 includes discarding events, calls, or user terminals having very abnormal behavior. The normalization operation further includes bringing the collected information into a standardized format. The normalized data is transmitted in a second step of the general procedure illustrated in FIG. 1 to a data store 116 or any other data storage device.

In the data warehouse 116, normalized data received from the data collection component 112 is aggregated in accordance with predetermined aggregation rules. The data aggregation operation is performed in the third step of the general procedure illustrated in FIG. Possible aggregation rules may, for example, specify that for a particular user terminal 98 the durations of individual communication sessions are summed (for example, in connection with a particular telephone service). Another aggregation rule may specify that for communications including a particular user terminal 98, all communications partners are retained.

As a result of the aggregation process, the data warehouse 116 provides a data structure representing user terminals 98 and relationships between terminals obtained from one or more communication parameters related to communications between user terminals, as illustrated in flowchart 400 of the method flowchart of FIG. 4 (step 402). The data structure or other component in the data store 116 or external to the data store 116 may also provide information about the use of the service for each of the many services used by the plurality of terminals 102 (see step 404 of the flowchart 400 of the method flow).

FIG. 2 illustrates an illustrative embodiment of a data structure 200 that results from aggregation of normalized data in the data warehouse 116 as received from the data collection component 112. In the example illustrated in FIG. 2, the data structure 200 comprises, for each user terminal 98, a plurality of terminals 102 (see FIG. 1) a separate data set. Each data set is identified using terminal ID 202. The terminal ID 202 may be a permanent network address of a user terminal, such as its mobile station ISDN number (MSISDN).

Each data set contains a data set part 204 indicating communication relationships between terminals of a particular user terminal 98. As shown in FIG. 2, communication relationships can be identified using the terminal IDs of frequent communication partners. In addition, each data set contains an additional data set part 206 that quantifies the individual communication relationships listed in part 204. Possible quantitative definitions of the communication relationship may include communication parameters, such as the total duration of communication with a particular communication partner, the number of communication sessions with a specific communication partner, the direction of communication, as well as the type of communication service used, or can be obtained from these parameters. Thus, the data set part 206 will also include information about the use of the service for various communication services used by the specific user terminal 98.

Alternatively, each data set may further comprise a specialized data set part 208 for storing information about the use of an additional service. Information about the use of additional services stored in part 208 of the data set, for example, may contain information regarding the use of application services or any other non-communication services.

The aggregation operation performed by the data store 116 may be performed in a time-controlled mode. For example, aggregation can be done on a per-day basis, on an hourly basis, or by dividing between days of the week and weekends. For each specific time unit (for example, hour, day, etc.), separate data sets can be generated in accordance with the underlying data structure.

Now, referring to the operation of the data analysis subsystem 106 and the additional step of the general procedure illustrated in FIG. 1, for each individual user terminal 98 (or subscription), a network graph for each specific service (for example, for each communication service) is obtained based on the data stored as a data structure 200 in the data warehouse 116. A network graph is generated using the network analysis component 118 of the data analysis subsystem 106. In short, the network analysis component 118 generates for each user terminal and service a graph similar to the network graph 300 illustrated in FIG.

The network graph 300 of FIG. 3 consists of nodes (such as nodes 302 and 304) and edges (such as edge 306) connecting every two nodes. The nodes of the network graph 300 are representatives of user terminals 98, while the edges are obtained from communication relationships between the terminals.

The illustrative network graph 300 of FIG. 3 is obtained based on the data set illustrated in FIG. 2 for the illustrative node 302 representing the user terminal with ID 123456789. As it becomes clear from the data set part 204 in FIG. 2, this user terminal has (direct ) communication relationships with seven other user terminals. For this reason, the node 302 is directly connected by edges to seven neighboring nodes (such as, for example, the node 304, which may correspond to the user terminal with ID 123456781). In order to reduce the complexity of the network graph 300, only communication relationships that exceed a predetermined lower threshold (for example, in terms of the number of communications or the total duration of communications) will be listed in part 204 of the data set.

It should be noted that after generating the network graph 300 using the network analysis component 118, not only the direct neighboring nodes of the node 302 are taken into account (for example, the direct communication partners listed in part 204 of the data set), but also not the immediate neighboring nodes (also called transitive neighboring nodes). As is understood in the present application, non-immediate neighboring nodes are immediate neighboring nodes of neighboring nodes of a node 302, which themselves are not immediate neighboring nodes of a node 302. Thus, several levels of non-immediate neighborhood can be determined. As shown in FIG. 3, the immediate neighboring nodes of the node 302 (such as the node 304) belong to the first level 310 of neighboring nodes, while the non-immediate neighboring nodes of the node 302 are located at additional levels 312, 314 of neighboring nodes, etc., depending on whether they are direct or non-direct neighboring nodes of direct neighboring nodes of node 302.

The number of levels of neighboring nodes accounted for for a specific node 302 in a particular network graph 300 is usually limited to reduce the overall complexity of the network graph 300 and its topological analysis. In many cases, it will be sufficient to take into account from one to three levels of neighboring nodes. As will be understood, the identification of neighboring nodes at external levels 312 and 314 of neighboring nodes of node 302 requires an assessment of the data sets that are provided in the data store 116 for neighboring nodes of node 302 (as, for example, for node 304).

Based on the various levels 310, 312, 314 of the neighboring nodes, a distance measure can be determined in the network graph 300. For example, the immediate neighboring nodes of the node 302 have the smallest distance from this node 302, the neighboring nodes of the first level (level 312 of neighboring nodes) will have the second smallest distance from the node 302, etc. As a result, a distance criterion can be defined for the graph-based service selection procedure, which will be described in more detail below.

In an optional embodiment, each edge (such as an edge 306 connecting two nodes 302 and 304) is additionally associated with metric information that quantifies the relationship (or relationship) between two nodes connected by an edge. Information about the metrics can be obtained from the communication parameters, as stored for a particular user terminal in part 206 of the data set corresponding to the data set (see figure 2). In addition, metric information can be obtained from the topology of the network graph 300, as will be explained below.

Separate metric information elements may be provided for each of the many services used by a particular terminal. In the example of a particular communication service, information about the metrics can be obtained based on the direction of communication (for example, one-way communication, two-way communication, etc.), the number of connections (for example, the number of SMS sent from one user terminal to another), the total duration of the communication between two user terminals, etc.

In the simplest case, information about metrics consists of a single number generated on the basis of an estimate of the set of communication parameters and / or topology of the network graph. It is important to note that separate information about the metrics can in principle be generated for each service, which allows you to determine the relationship between users (including IM services, group-based communication services, certain application services, etc.).

In addition to generating the network graph 300, the network analysis component 118 is also responsible for analyzing the generated network graph 300 in connection with various topological criteria that can later be used in connection with the service selection procedure. Such topological criteria may include the belonging of individual nodes to the so-called groupings, the presence of transitive relations (i.e., not immediate neighboring nodes), etc. In addition, each node in the network 300 may be ranked according to topological criteria, and then the ranking may be taken into account during the service selection procedure. The operations of analyzing the network graph and ranking nodes as performed by component 118 will be described in more detail later with reference to FIG. 5 of FIG. 11.

In the fifth operational step of the general procedure illustrated in FIG. 1, service usage information, as provided by the data store 116, is analyzed in connection with individual services and a specific user terminal 98. As shown in FIG. 1, a service utilization analysis is performed with using a specialized component 120 of the use of the service. It should be noted that the analyzed services are not necessarily limited to the services actually used by the user terminal under study (such as the user terminal corresponding to the node 302 in FIG. 3). Instead, we will analyze the use of the service for all nodes in a particular network graph (as, for example, for node 304). As a result of the analysis of the use of the service, a list of services used in this network column 300 of FIG. 3 will be determined.

In an optional mode, an analysis of the use of the service further includes ranking the service. Each analyzed service can be ranked, for example, in accordance with the full use of this service among nodes belonging to a particular network graph. In other words, the service with the highest full use of the service (for example, in terms of duration of use or number of uses) will be ranked highest. Service ranking can be performed separately for each level of neighboring nodes. In the network graph scenario illustrated in FIG. 3, this approach would thus result in three separate service rankings.

In the next operational step of the general processing procedure illustrated in FIG. 1, for some or more separate user terminals 98, coverage component 122 analyzes (or “covers”) the network graph 300 and associated information (about metrics and other) as generated by analysis component 118 network, in combination with the ranking (s) of the service, as defined by component 120, which is responsible for the analysis of the use of the service. The purpose of this analysis, performed by the overlay component 122, is to select one or more ranked services for recommendation for a particular user terminal corresponding to the node 302 in the network column 300.

Specifically, the selection is made in such a way that a highly ranked service is currently not being used by a particular user terminal in order to increase the connectivity and / or usability of a particular user terminal. As illustrated in block 406 of the flowchart 400 of the method (see FIG. 4), a service selection procedure is based on service usage information obtained for neighboring user terminals having a predetermined association with a particular user terminal, for example, from a point view of the network graph 300.

During one possible implementation of the service selection procedure, a distance measure is applied according to which the likelihood of a particular selected service increases if many nearby neighboring nodes use the service (such as neighboring nodes within the first level 310 of neighboring nodes illustrated in FIG. 3). The basis for this approach for analyzing associated neighboring user terminals is the fact that the connectivity and / or usability of a particular user terminal, in particular, would increase if the selected service was already used with the help of nearby neighboring nodes. In other words, if a service is recommended to user terminal A, because user terminals C, D, and E in the network graph of user terminal A use this service, this recommendation is cross-checked as to how close the user terminal A and user terminals C are in the network graph, D and E. In the case where the accumulated distance measures for the three user terminals C, D and E in connection with the user terminal A (or any other distance criteria) would exceed a predefined threshold, a specific service will no longer be considered.

During the service selection procedure performed by coverage component 122, various individual criteria and combinations of criteria may be considered. As mentioned above, the rank of an individual service can be taken into account, as well as the proximity of user terminals already using this service. In addition, metric information associated with each edge in the network graph 300 can be taken into account, as well as optional rankings of the various nodes in the network graph 300. Other topological criteria of the network graph 300 can also be considered.

Coverage component 122 outputs, for each analyzed user terminal 98, indications of one or more services sent to the corresponding user terminal 98 via the communication network. The service instructions output by the coverage component 122 are sent in the seventh operation step to the optional service recommendation component 124, which forwards these service recommendations in the eighth operation step to the messaging subsystem 108, as shown in FIG.

Service recommendation component 124 identifies one or more microsegments in a network graph 300 of a particular user terminal (node 302) based on an analysis performed by coverage component 122, service utilization component 120, and network analysis component 118. Specifically, the microsegments identified by the service recommendation component 124 include nodes in the network graph 300 having similar service usage behavior and / or topological relationship. In identified microsegments, the service can then be recommended based on the level of service penetration in the nodes making up the microsegment. In other words, if many nodes in the identified microsegment use a specific service, then this service can be recommended to other nodes in the microsegment that have not yet used this service. The identification of microsegments in the network graph 300 can be performed using the methods described below with reference to FIGS. 5 to 11.

In the messaging subsystem 108, the service indications related to the services selected by the coverage component 122 (and possibly the filtered service recommendation component 124) are received together with the indication of the particular user terminal 98 for which the service was selected using the message means 126. Then, the message means 126 contacts in the ninth step of the processing chain with a database or repository (such as a location reference register or HLR) to obtain contact information (e.g., a fixed network address) of a user terminal 98 for which a service has been selected. Then, in the tenth step illustrated in FIG. 1, a message (eg, SMS or email) indicating one or more selected services is sent to the user terminal 98 in the aggregate of 102 terminals using its network address as received from the subscription repository 128. A message is sent via the output interface 130 of the messaging subsystem 108 and informs the user terminal 98 of the availability (e.g., availability) of one or more selected services. This process corresponds to step 408 in flowchart 400 of the flowchart of FIG. 4. The user terminal 98 receiving the message may be configured to automatically initiate all necessary steps to gain access to the recommended service. Alternatively, user interaction (such as user confirmation) may be requested explicitly.

In addition to displaying a service indication message, a flag or other identifier is set in the subscription repository 128. The flag setting is then evaluated at the final eleventh step after receiving feedback from a specific user terminal 98. Feedback may be in a message requesting, for example, downloading software, updating a subscription, etc., in order to make a recommended service available for a particular user terminal 98. By evaluating the feedback based on the flag settings, appropriate feedback can be provided from the messaging subsystem 108 to the subsystem data analysis topic 106 to implement machine comprehension algorithms in the service selection procedure applied by coverage component 122.

The service indication message sent from the message means 126 to the user terminal 98 may specifically refer to an available service and / or describe an available service. A service reference in a message may include, for example, an identifier, an IP address, or a universal resource address (URL) of a service. Alternatively, the message may also include information regarding the steps to be taken in order to enable the use of a particular service. In addition, the message may include an indication of user terminals that are close, in terms of network graph 300, to a recommendation recipient 98 who already uses this service. In addition, the service utilization profile in connection with the recipient 98 may be enriched with information collected for user terminals that are close to the recipient in terms of network graph 300.

Thus, having described the general procedure illustrated in FIG. 1 in some detail, some optional modes of operation of the network analysis component 118 regarding network topology analysis will now be further discussed. The result of the topological analysis can be used to identify criteria for the service selection process performed by coverage component 122.

Interpreting the network graph 300 of FIG. 3, as (also) indicating to a certain extent a social network, it is obvious that the indication or recommendation of the service that is performed for the service used by the user terminal managed by a close friend of the operator of the user terminal under investigation will be especially strong and useful for last one. For this reason, there is a need for an abstract mechanism to determine the “strength” of a particular relationship between terminals (or between users) (ie, the “strength” of a connection corresponding to an edge in network graph 300).

In the future, several abstract mechanisms will be discussed in order to deal with this need only on the basis of an analysis of the topology of the network graph. The resulting mechanisms can be implemented by the analysis component 118 of FIG. 1 in the context of providing additional metrics information in relation to nodes and / or edges. Then, this additional metric information may form an additional basis for the service selection operation performed by coverage component 122. However, it will be understood that the methods discussed later in this application can also be performed in any other contexts. In particular, the following methods can be performed in the context of recommending elements other than services to objects other than user terminals.

From the point of view of the network graph 300, social groups will appear as the so-called clusters of individual nodes. One typical feature of clusters is the fact that they are closely connected by edges representing communication lines. If the cluster is fully connected, all nodes (also called vertices) in the cluster are connected to each other, they are as closely formed in the cluster as possible.

Groupings

A complete graph is a graph in which all nodes (sometimes also called vertices) have direct connections (for example, edges) to each other. In a larger graph, it is interesting to find subgroups of nodes that form a complete graph. A graph can contain several subgroups that are complete, and figuring out how many there are, the largest subgroup and what they are, in graph theory is called the grouping problem (finding the maximum of the groups). Figure 5 depicts a grouping that is created by four nodes {V0; V1; V2; V3}. These four nodes are all fully connected, i.e. all nodes have edges with each other and form the largest grouping in the network graph 500 of figure 5 (grouping four). Unfortunately, the grouping problem is non-deterministic polynomial time (NP). This means that it is difficult, if not impossible, to create an algorithm that can be easily scaled, and the computational time increases exponentially with the size of the graph. Thus, a workaround or approximation of the grouping problem would be beneficial.

All groupings larger than three can be divided into groups of size three (i.e., having equally three nodes each). This means that all groups can be represented as a set of groupings of size three. A formula that calculates how many subgroups exist in a large grouping of size n is determined using the following formula:

.

.

, группировок размера три. Пользовательский терминал в сетевом графе с двумя группировками, одной размера 3 и одной размера 4, имел бы объединенную группировку размера 9. Таким образом, можно сравнивать разные пользовательские терминалы относительно того, сколько группировок, и как велики их группировки. Обнаружено, что число группировок, которые обычно имеют два пользовательских терминала, будет мерой того, насколько хорошим является их отношение. Таким образом, соответствующая информация о метриках может быть добавлена к их отношению (т.е. к взаимно соединяющему ребру на рисунке сетевого графа) и оценена во время процедуры выбора услуги.For example, a grouping of size n = 4, as illustrated in FIG. 5, could be divided into 6,

, groupings of size three. A user terminal in a network graph with two groups, one size 3 and one size 4, would have a combined grouping of size 9. Thus, it is possible to compare different user terminals with respect to how many groups and how large their groups are. It has been found that the number of groupings, which usually have two user terminals, will be a measure of how good their relationship is. Thus, the relevant metrics information can be added to their relationship (i.e., to the interconnecting edge in the network graph) and evaluated during the service selection procedure.

Transitive relationship

Transitive relationships are relationships that have a distance in the graph from the node in question (such as node 302 in FIG. 3), which is greater than one, i.e. which correspond to non-immediate neighboring nodes in the terminology used above. Assume that nodes that have a relationship with the node in question through another node do not have a close connection with this node, like nodes that are directly adjacent nodes.

When the distance increases, the strength of the relationship decreases, therefore, it is appropriate to limit the depth of how far to seek the relationship. The number of “mutual friends” will also influence the outcome of the relationship. In this case, the transitive relation is stronger than when there is only one common friend. The network graph 600 illustrated in FIG. 6 compares two nodes V4, V5, both of which have a distance of two from the active node, but consider that one of the nodes has a stronger relationship due to the number of “common friends” that he has with active node V0.

Calculation of the inference relation value

для непосредственного соседнего узла

задают формулой, приведенной ниже. Следует заметить, что она не применяется для узлов, которые имеют расстояние, большее единицы, от рассматриваемого узлаLogically inferred relationship, or rank (user rank) of the node in question

for the immediate neighboring node

set by the formula below. It should be noted that it does not apply to nodes that have a distance greater than one from the node in question

- число совместных группировок, элементами которых являются узел

и узел ν , а

- полное число группировок, элементом которых является узел ν .

- максимальная глубина, которую принимает во внимание алгоритм, узлы, которые имеют большее расстояние, чем это, не получат логически выведенного значения для рассматриваемого узла. Деление на полное число группировок выполняют, чтобы дать процентное отношение числа группировок, элементом которых является узел. Более высокое процентное отношение означает более высокий ранг.Where

- the number of joint groupings whose elements are a node

and the node ν , and

- the total number of groups whose element is the node ν .

- the maximum depth that the algorithm takes into account, nodes that have a greater distance than this will not receive a logical value for the node in question. The division by the total number of groups is performed to give a percentage of the number of groups of which the node is an element. A higher percentage means a higher rank.

The rank calculation for transitive node relationships is different from the rank calculation for immediate neighboring nodes. Nodes that do not have direct links to the node in question will have a lesser relationship. However, a greater number of “mutual friends” will result in a higher rank. Therefore, the ranking will partially depend on how many neighboring nodes of the node in question are connected to a particular node.

Since different neighboring nodes of the node in question have different ranks, they can be weighted when calculating the transitive relation. The node under consideration is more likely to have a strong connection (in terms of metric information) with a specific node if they have a strong connection with a “common friend”.

The rank of a particular site can be calculated using the formula below:

является константой, которая определяет баланс между важностью числа общих друзей и значением ранга соседних узлов. Абсолютные или относительные ранги, ассоциированные с соседними узлами рассматриваемого узла, могут быть оценены в контексте с процедурой выбора услуги. Иначе говоря, вероятность выбора конкретной услуги, используемой непосредственным соседним узлом, имеющим более высокий ранг, выше по сравнению с другой услугой, используемой прямым узлом, имеющим более низкий ранг. Таким образом, можно гарантировать, что будут рекомендованы услуги конкретной используемости.In the above formula, the rank increases when the number of “common friends” increases, and additionally increases if the values of these edges are large. It should be noted that the parameter

is a constant that determines the balance between the importance of the number of common friends and the rank value of neighboring nodes. The absolute or relative ranks associated with the neighboring nodes of the node in question can be evaluated in the context of the service selection procedure. In other words, the probability of choosing a particular service used by a direct neighboring node having a higher rank is higher compared to another service used by a direct node having a lower rank. In this way, it can be guaranteed that the services of a particular usability are recommended.

Approximation

In the following, a second approach is presented to calculate the ranks of direct or immediate neighboring nodes. This approach includes the approximation of group counting and partitioning of groups. The idea is to look for groupings of the node in question in order to predict the relationship. By splitting the groupings as described above, one can find ever smaller groups. Of these, only divided groups that have the node under consideration as an element will be additionally considered.

The network graph 700 of FIG. 7 depicts a scenario in which groupings are calculated based on the node V0 in question as an element of groupings. The node V6 is an element of three groupings ({V6, V7, V0} {V6, V4, V0} {V6, V5, V0}), all of them are three times the size.

If the grouping {V6, V6, V6, V6} is divided into smaller groups, all groups should still have V0 as an element, so it would be divided into {V0, V1, V2}, {V0, V1, V3 } and {V0, V2, V3} (it should be noted that {V1, V2, V3} do not have V0 as an element and, therefore, do not use it). You can see that each of the nodes is an element of two groups, with the exception of V0. This means that in this division of the group there is a loss of one group. This can lead to a different result between the usual group counting algorithms, but it will give a good approximation.

Since the number of groups whose elements are nodes is related to the total number of input and output edges of the node, the number of groups can be easily calculated by subtracting a unit from the number of input and output edges. For example, V2 has 3 edges and, therefore, is an element of 2 groups, the element of which is also V0.

Конкретно будут использованы следующие определения:You can calculate the rank of a particular node using the same formula as in the previous section, but with different definitions

Specifically, the following definitions will be used:

The trust

When trust relationships between users of different user terminals are abstractly defined, for example, using topological features of such groupings and transitive relationships, as summarized above, the network graph 800 can be visualized, as shown in Fig. 8. Directional edges are used in this network graph 800, given that “trust” is always transitive and asymmetric.

In the case of a directed graph, as shown in Fig. 8, a special definition of grouping is applied. This definition of grouping is that if the node in question has two or more neighboring nodes that are “trusted”, and if they are connected in any direction of the graph, then this graph forms a grouping. Fig.9 depicts an example of such a grouping, in which the considered node is V0.

Performance check

A performance test was performed to compare a calculation algorithm that separates groupings (called group counting) and an approximation algorithm that counts edges (called edge count). The results show that the edge counting algorithm is faster than the grouping counting algorithm in terms of performance. The difference increases when the average cluster size in the network graph increases. This is due to the fact that there are more edges when there are large clusters, and therefore more computations, which proves its scalability (see diagrams 1000 and 1100 of FIG. 10 and 11, respectively).

Recommendations - predicting element values

In order to be able to recommend items (such as services) there must be a way to predict how much a user terminal (or any other entity, such as a user) would benefit from a particular item. Winning can be predicted using the following formula:

where c (i) is the estimate predicted for element i , sim (c, p) is the similarity between user terminals (or other objects) c and p , and p (i) is the nominal value of the user terminal (or object) p relative to the element i . Similarities between objects are in this case the value of the relationship between objects, as determined using any of the approaches discussed above. Using this formula, he can see that objects with a higher likeness have a higher effect on the value of an element. A higher element value indicates a better recommendation for c . This is done for all elements that have been evaluated by immediate neighboring objects. Elements with the highest values will be recommended to the object in question.

An advantage of the above method, as illustrated in FIG. 5 of FIG. 11, is the fact that it checks patterns in a graph that indicate close relationships compared to conventional (personalized) trusted systems. Since it only checks patterns, it is not limited to any data, which makes it very suitable for all kinds of network graphs and, in particular, for network graphs that have directed edges (i.e. in which trust can be extracted). A template-based method can be applied (for example, by component 124 of the service recommendation of FIG. 1) to identify microsegments in column 300 from nodes exhibiting similar behavior, and to filter those services from which the user terminal in question could not benefit .

The interaction between the coverage component 122 and the service recommendation component 124 of FIG. 1 can be described as follows. In one implementation, coverage component 122 mainly determines the difference in service usage between user terminals in a given set of terminals, and then segments the set of terminals into groups (based on network graphs, as described above). In groups, “outsiders” (that is, user terminals in a group that do not use a particular service) are advised to use a specific service. Service recommendation component 124 can then further weigh service recommendations (e.g., filter out potentially inappropriate recommendations), as discussed above with reference to FIG. 5 of FIG. 11.

As it became clear from the above, the methods of indicating services described in this application, in particular, are useful in increasing the connectivity and / or use of individual user terminals when applying complex procedures for selecting services. By taking into account the relationship between the terminals in the procedure for selecting a service, indications of the availability of services can be made “personal”, so that the user can better understand the benefits of a particular recommendation of the service. This advantage is the result of possible cross-correlations between available services, on the one hand, and terminal-specific network graphs, on the other. In addition, the methods described in this application can provide explicit feedback regarding the recommendations of services and requests for services, which makes it possible to gradually refine the procedure for choosing a service.

Although the present invention has been described with respect to specific illustrative embodiments, those skilled in the art will understand that the present invention is not limited to the specific examples described and illustrated in this application. It should be understood that this disclosure is illustrative only. Thus, it is intended that the invention be limited only by the scope of the claims appended hereto.

Claims (19)

1. A method of providing an indication of service availability to a user terminal (98), comprising the steps of:
provide a data structure (200) representing user terminals and relationships between terminals obtained from one or more communication parameters related to communications between user terminals,
provide at least some of the user terminals with information about the use of the service for each of the many services,
for at least one user terminal (98) and from a variety of services, a service not used by at least one user terminal (98) is selected, the service selection being based on service usage information provided for user terminals having a predetermined association with at least one user terminal (98), as defined by relations between the terminals, and
sending a message indicating the selected service to at least one user terminal (98). 2. The method according to claim 1, in which the data structure (200) can be represented as a network graph (300) with nodes (302, 304) and edges (306) connecting every two nodes (302, 304), and in which nodes (302, 304) correspond to user terminals, and edges (306) are obtained from relations between the terminals. 3. The method according to claim 2, in which the predefined association underlying the choice of service contains a criterion expressed as a measure of distance applicable to the network graph (300). 4. The method according to claim 2 or 3, in which each edge (306) is associated with information about the metrics, quantifying the mutual relationship between two nodes (302, 304) connected by the edge (306), and in which the information about the metrics is obtained from one or more communication parameters. 5. The method according to claim 4, in which the predefined association underlying the choice of service contains a criterion applicable to metric information. 6. The method according to claim 2 or 3, in which one or more ribs (306) are directed, and in which the stage at which the service is selected, contains a criterion applicable to the direction of the ribs. 7. The method according to claim 2 or 3, in which at least some of the user terminals are ranked based on at least one of: group membership and the presence of transitive relationships from the point of view of the network graph (300), and in which step, which selects the service, contains a criterion expressed in terms of ranking. 8. The method according to claim 2 or 3, in which each edge (306) is associated with separate information about the metrics for each of the many communication services. 9. The method according to claim 2 or 3, in which information about the use of the service relates to communication services, and in which the selected service is a communication service. 10. The method of claim 9, wherein the communication services include one or more of voice services, text services, data services, multimedia services, instant messaging services, and group services. 11. The method according to claim 2 or 3, in which information about the use of the service relates to application services, and in which the selected service is an application service. 12. The method according to claim 2 or 3, further comprising the step of ranking the services in accordance with information about using the service, and in which the step of selecting the service is further based on the ranking of the service. 13. The method of claim 12, wherein the ranking of the service depends on the full use of each individual service by a plurality of user terminals. 14. The method according to claim 2 or 3, in which one or more communication parameters include at least one of: communication partner or partners, communication duration, number of communications, communication direction and type of communication service used. 15. The method according to claim 2 or 3, in which one or more communication parameters can be obtained from at least one of: call detail records, event detail records, research of a communication network interface and user terminal agents (110). 16. A computer-readable storage medium containing instructions executed by a computing device to cause the computing device to perform the steps of the method according to any one of claims 1 to 15. 17. An apparatus for providing an indication of service availability to a user terminal (98) comprising
a memory (116) providing a data structure (200) representing user terminals and relations between terminals obtained from one or more communication parameters related to communications between user terminals, and further providing at least some of the user terminals information about using the service for each of the many services
a processor (122) adapted to select for at least one user terminal (98) and from among a plurality of services a service not used by at least one user terminal (98), wherein the choice of service is based on service usage information provided for user terminals, having a predefined association with at least one user terminal (98), as defined by the relationship between the terminals, and
an output interface (130) for sending a message indicating the selected service to the user terminal (98). 18. The device according to 17, in which the data structure (200) is configured to be represented as a network graph (300) with nodes (302, 304) and edges (306), in which nodes (302, 304) are representatives of user terminals, and in which the ribs (306) connect the two nodes (302, 304), and they are obtained from the relationship between the terminals. 19. The device according to p, in which the predefined association underlying the choice of services, contains a criterion, expressed as a measure of distance, applicable to the network graph (300).

Images