[go: up one dir, main page]

WO2015178811A1 - Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire - Google Patents

Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire Download PDF

Info

Publication number
WO2015178811A1
WO2015178811A1 PCT/SE2014/050629 SE2014050629W WO2015178811A1 WO 2015178811 A1 WO2015178811 A1 WO 2015178811A1 SE 2014050629 W SE2014050629 W SE 2014050629W WO 2015178811 A1 WO2015178811 A1 WO 2015178811A1
Authority
WO
WIPO (PCT)
Prior art keywords
time intervals
time interval
value
network device
value vector
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/SE2014/050629
Other languages
English (en)
Inventor
Rickard CÖSTER
Vincent Huang
Zhili WU
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Telefonaktiebolaget LM Ericsson AB
Original Assignee
Telefonaktiebolaget LM Ericsson AB
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Telefonaktiebolaget LM Ericsson AB filed Critical Telefonaktiebolaget LM Ericsson AB
Priority to PCT/SE2014/050629 priority Critical patent/WO2015178811A1/fr
Publication of WO2015178811A1 publication Critical patent/WO2015178811A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Definitions

  • the invention relates to a method, network device, computer program and computer program product for identifying a user behaviour pattern in a cellular system.
  • Association rule mining can be used to find relevant services a user uses together with other services and can be used to predict user behaviour as well.
  • Knowing behaviour patterns is useful for optimization and dynamically provisioning resources in a cellular system, and also how an operator can handle individual customers in a cellular system.
  • 2014/035305 describes a data collector arranged to collect data regarding application usage in an end user device.
  • International patent application WO 2012/ 064278 describes matching of a first location profile with one other location profile.
  • a method for identifying a user behaviour pattern in a cellular system the method being performed by a network device of the cellular system.
  • the method comprises the steps of collecting data of services a user of the cellular system has used and at what time, and dividing the collected data of services into time intervals.
  • the method further comprises the steps of defining at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services, and comparing the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof.
  • the method comprises the steps of combining the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals, and comparing the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof.
  • the method comprises repeating the steps of combining and comparing the value vector of each time interval of the reduced time intervals, until a value vector is combined into a single time interval.
  • the method may further comprise the step of clustering a plurality of users of the cellular system together, by comparing their respective user behaviour pattern. Further optimization of resources, such as bandwidth, in the cellular system is provided in this way.
  • resources such as bandwidth
  • the step of combining may comprise uniting the vector value of each time interval of grouped adjacent time intervals, to facilitate clustering of a plurality of users of the cellular system together.
  • the step of combining may comprise averaging the vector value of each time interval, or re-mining the reduced number of time intervals, wherein a new set of rules is created for a reduced set of number of time intervals.
  • Each value vector may include one or more of support, confidence and lift values, to define a stable set of rules in the time intervals.
  • the time intervals may be made up by one or more of the following: an hour of a day, a day of a week, a month of a year, a public holiday, and a set of one or more of hour, day, and month, to facilitate clustering of users.
  • At least one of the time intervals may comprise at least two different data of services
  • the step of defining may comprise defining a value vector for each data of service for each time interval of the time intervals, by association rule mining the collected data of services.
  • a network device configured to identify a user behaviour pattern in a cellular system.
  • the network device comprises a processor and a computer program product storing instructions that, when executed by the processor, causes the network device to collect data of services a user of the cellular system has used and at what time.
  • the network device further comprises instructions to divide the collected data of services into time intervals and to define at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services.
  • the network device also comprises instructions to compare the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof and to combine the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals. Further, the network device comprises instruction to compare the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof. Finally, the network device comprises instruction to repeat the steps of combine and compare the value vector of each time interval of the reduced time intervals, until a value vector is combined into a single time interval.
  • a network device configured to identify a user behaviour pattern in a cellular system.
  • the network device comprises a collect manager, a define manager and a combine manger.
  • the collect manager is configured to collect data of services a user of the cellular system has used and at what time.
  • the define manager is configured to divide the collected data of services into time intervals, and to define at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services.
  • the combine manager is configured compare the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof, and to combine the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals.
  • the combine manager is further configured to compare the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof, and to repeat combine and compare the value vector of each time interval of the reduced time intervals, until a value vector is combined into a single time interval or a value vector distance between the time intervals is greater than a distance threshold value.
  • the computer program comprises computer program code which, when run on a network device of a cellular system, causes the network device to collect data of services a user of the cellular system has used and at what time and to divide the collected data of services into time intervals.
  • the program code further causes the network device to define at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services and to compare the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof.
  • the program code also causes the network device to combine the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals and to compare the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof. Finally, the program code causes the network device to repeat the steps of combine and compare the value vector of each time interval of the reduced time intervals, until a value vector is combined into a single time interval. Thereby a user behaviour pattern can be identified by how the time intervals are divided and the value vector in each time interval of the time intervals.
  • a computer program product comprising a computer program and a computer readable storage means on which the computer program is stored.
  • Fig. 1 is a schematic diagram illustrating an environment
  • Figs. 2a and 2b are schematic diagrams illustrating identified user patterns divided into time intervals over a day;
  • Fig. 3 is a schematic diagram illustrating rule sequences and comparisons there between;
  • Figs. 4A-4B are flow charts illustrating methods for embodiments presented herein;
  • Fig. 5 is a schematic diagram illustrating some components/devices of a network device
  • Figs. 6A-6B are schematic diagrams illustrating various locations where the network device of Fig. 5 can be implemented;
  • Fig. 7 is a schematic diagram illustrating a user characterization
  • Fig. 8 is a schematic diagram showing functional modules of a network device; and Fig. 9 is a schematic diagram illustrating a Hadoop system.
  • Different rules may apply for different time intervals. Also, different users may behave the same or similar with shifted time. One user may wake up six in the morning and another may wake up eight in the morning, but they may do similar tasks after they get up.
  • Users may also use similar services during a day, in different order. One may read news and play games in the morning; the other may use very similar services in the evening. User behaviour in different time interval can be found and also similar users can be found despite of their behaviour is shifted or even in different order in time.
  • Day parting is used in broadcasting programming.
  • a weekday is often manually divided into several fixed parts, like early morning (4:00-7:00 AM), breakfast (7:00-9:00 AM), Daytime (9:00 AM-5:oo PM), school newscast time (6:30-7:00 PM), early evening (7:00-8:00 PM), national prime time (8:00-11:00 PM), late night (11:35 PM-2:oo AM) and graveyard slot (2:00-4:00 AM).
  • fixed day parts are no longer relevant due to different behavioural patterns of users
  • a stable set of rules is meant a set of rules stable during the time intervals
  • An approach is to find personalized time intervals.
  • Each of the personalized time intervals contains a set of stable rules that defmes what services a user is using during a time interval.
  • the rule sets may be described with a set of features. From the set of stable rules per user, known techniques such as Dynamic time-warping may be used to find similar users.
  • usage patterns can be used to group users in various ways, such as to group together all users that have most of their usage during peak hours, and then by correlating with other network data further divide this group into e.g. low/medium/high signalling users. Users causing a lot of signalling during peak hours, as their normal behaviour, may thus be identified as doing so also in the future.
  • changed user behaviour may be a sign of the account being used by others, which may be used for fraud detection.
  • User data is collected from the cellular system.
  • the user data contains what services each user has used and at what time.
  • User behavior characterization processes the user data and identifies distinct periods which different rule sets can be applied to.
  • a user similarity identifies similar users based on how the rule set changes for each user.
  • Item sets are formed from e.g. the mobile apps used by a user in each time interval, for example an hour, of a day. Each day then has a transaction for a particular time interval.
  • FIG. 7 illustrates how user characterization of user data may be applied.
  • a transaction Ti has an activity A, which is followed by a transaction T2 having an activity B, which is followed by a transaction T3 having an activity C, which is followed by a transaction T4 again having an activity A, which is followed by a transaction T5 having an activity D.
  • a user may have the following transactions:
  • association rule mining for example the apriori algorithm and its variants
  • rules are identified giving both so called Support and Confidence over the items in the same time interval over a test period of time.
  • the support supp(X) of an itemset X is defined as the proportion of transactions in the data set which contain the itemset.
  • supp(X U Y) means "support for occurrences of transactions where X and Y both appear", not "support for occurrences of transactions where either X or
  • suppO is a set of preconditions, and thus becomes more restrictive as it grows (instead of more inclusive).
  • Confidence can be interpreted as an estim ate of the probability P(Y
  • the conviction can be interpreted as the ratio of the expected frequency that X occurs without Y (that is to say, the frequency that the rule makes an incorrect prediction) if X and Y were independent divided by the observed frequency of incorrect predictions.
  • Each rule is associated with a vector of values, including Support, Confidence, Lift, etc. that characterizes how well the rule is supported by the data.
  • Each time interval is then associated with a numeric rule-vector of multiple rules.
  • Adjacent time intervals are next joined together.
  • the simplest approach is to join two neighboring time intervals if they have the minimum distance among all distances of eligible pair of hours. Distance in this case is a measure of how similar the service usage is and how well supported this service usage is by the data, within the two adjacent time intervals.
  • FIG. 2a An example of joined adjacent time intervals is illustrated in Fig. 2a.
  • Each hour from 1 to 24 in Fig. 2a has a transaction, and there are thus 24 time intervals.
  • hours 1-5 are joined into a new time interval and hours 22-24 are joined into a new time interval.
  • Each of the hours 1-5 and 22- 24 thus have a minimum distance to a neighboring hour compared to all hours 1-24.
  • hour 6 is joined to hours 1-5 and hours 16 and 17 are also joined together.
  • the eligible number of time intervals after the second round is thus 16 time intervals.
  • all time intervals are joined into a single time interval.
  • the value vector of the rule in the combined time interval is the average of the value vectors in Ti and T2.
  • Re-mining After combining Ti and T2, all previous mined rules are discarded and new rules are mined for the combined time interval.
  • Uniting and averaging may be used together.
  • the rules may be set as consecutive rules of item sets.
  • sequential rules are mined. Incremental methods may be devised too for sequential rule generation. Mining association rules from a full data set every time new data becomes available is not always feasible since the size of data may be very large.
  • the interval-joining procedure maybe conducted recursively, until all intervals are combined into a single interval or until the distance between two time intervals is too great, over a vector distance threshold.
  • Calculations for the interval-joining procedure maybe made by computer processing, and each interval-joining maybe stored to be able to extract different levels of dissimilarities of each user behavior pattern.
  • a (tree) dendrogram may thus be generated, as an illustration of the interval-joining procedure but is not necessarily generated as such, for all dayparts, or at least of desired dayparts. Different granularity of dayparts can then be extracted depending on different application scenarios, since the whole tree is depicted of desired dayparts. Dayparts of a 24 hour time interval may e.g.
  • the interval-joining process may be stopped. With knowledge of a desire to only extract dissimilarities below the vector distance threshold, further processing is not needed.
  • Fig. 2b illustrates another example of joined adjacent time intervals is forming a tree.
  • Each hour from 1 to 24 in Fig. 2a has a transaction, and there are thus 24 time intervals.
  • hours 8 and 9 are joined into a new time interval. Only the hours 8 and 9 thus have a minimum distance to a neighboring hour compared to all hours 1-24. After the first round there are thus now only 23 time intervals.
  • hours 2-5 are joined into a new time interval and hour 7 is joined to the joined time interval 8-9 hours.
  • the eligible number of time intervals after the second round is thus 19 time intervals.
  • all time intervals are joined into a single time interval.
  • the degree of dissimilarity of activities of a user is also illustrated in Figs. 2a and 2b.
  • the dendrogram will have a high degree of dissimilarity.
  • distance between two trees there are many ways to calculate distance between two trees. However, since the trees are partitioned according to the value vectors of mined rules, one approach is to calculate the Euclidean distance of the value vectors for the respective time intervals. If a rule does not appear in a time interval, a default value, e.g. o, may be used.
  • Mined user data of a first user may be compared to mined user data of a second user.
  • User similarity identification may be performed by e.g.
  • a sequence matching algorithm may be used, or direct compare of the tree structures may be used.
  • Fig. 3 illustrates activities A, B, C, D, E and inactive periods (I) for three users Ui, u 2 , and u 3 .
  • u 2 is matched against Ui and u 3 .
  • the arrows in Fig. 3 indicate the matching sequence elements. For example, u 2 against Ui matches by CAC versus CAIC, disregarding the inactive period I.
  • the inactive period can be treated as a wildcard, meaning that it can match zero or more activities in other sequences.
  • sequence matching algorithms already exist today, and this sequence alignment problem is similar to a genome sequence matching problem.
  • Each element A-E may not be a number, but a rule.
  • Each rule may be converted to a weighted vector of activities, wherein the weights define the strength of the activities during the time period.
  • the rules may be compared with each other using vector distance or similarity measures, and in this way different time periods are compared.
  • Bob u 2
  • Alice drives to work and Bob takes subway to work. Every morning, Alice checks weather forecast and traffic conditions. While Alice is driving, she uses navigation services. During lunch time, she reads news and Facebook updates. In the evening, she checks traffic before she goes home and uses navigation while she is driving. Bob takes subway in the morning. He checks subway schedules before he leaves home.
  • Identification of user behavior patterns may also be of use is in M2M systems, wherein algorithms control the regularity of service usage.
  • Structurally similar machines or devices may be identified based on the service usage, and for example allocate resources such as bandwidth based on this information.
  • the level of dissimilarity maybe selected to be about 5.
  • Fig. 1 is a schematic diagram illustrating a cellular system 4 providing an environment where embodiments presented herein can be applied.
  • the cellular system 4 is connected to a wireless communication device (WCD) 1 in connectivity with a base station 2, such as an eNodeB in a Long Term
  • the term wireless communication device may be or alternatively be termed as a mobile communication terminal, user equipment, mobile terminal, user terminal, user agent, machine-to-machine device etc., and can be, for example, what today are commonly known as a smartphone or a tablet/laptop with wireless connectivity.
  • the WCD 1 may, but does not need to, be associated with a particular end user.
  • the WCD 1 may also be a telematics unit embedded in a vehicle such as a car, bus and truck, and be connected to a vehicle-internal network for exchange of e.g. vehicle or driver data with a fleet management system connected to the vehicle via the cellular system 4.
  • the WCD 1 may also be a unit mounted in a dashboard of a vehicle for displaying information and communicating with the driver or passengers of the vehicle and being connected to the telematics unit embedded in the vehicle.
  • a network device 20 for a cellular system 4 is presented with reference to Fig. 5, which network device 20 is arranged to identify a user behaviour pattern in the network system 4.
  • the network device 20 comprises: a processor 30; and a computer program product 62 storing a computer program 64 with instructions that, when executed by the processor 30, causes the network device 20 to: collect 40 data of services a user of the cellular system has used and at what time; divide 42 the collected data of services into time intervals; define 42 at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services; compare 43 the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof; combine 44 the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals; compare 45 the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent
  • Fig. 5 is a schematic diagram showing some components of the network device 20.
  • the processor 30 maybe provided using any combination of one l6 or more of a suitable central processing unit (CPU), multiprocessor, microcontroller, digital signal processor (DSP), application specific integrated circuit etc., capable of executing software instructions of a computer program 64 stored in a memory.
  • the memory can thus be considered to be or form part of the computer program product 62.
  • the processor 30 maybe configured to execute methods described herein with reference to Figs. 4A-4B.
  • the memory may be any combination of read and write memory (RAM) and read only memory (ROM).
  • the memory may also comprise persistent storage, which, for example, can be any single one or combination of magnetic memory, optical memory, solid state memory or even remotely mounted memory.
  • a second computer program product in the form of a data memory 63 may also be provided, e.g. for reading and/or storing data during execution of software instructions in the processor 30.
  • the data memory 63 can be any combination of read and write memory (RAM) and read only memory (ROM) and may also comprise persistent storage, which, for example, can be any single one or combination of magnetic memory, optical memory, solid state memory or even remotely mounted memory.
  • the data memory 63 may e.g. hold other software instructions 65, to improve functionality for the network device 20.
  • the network device 20 may further comprise an I/O interface 61 including e.g. a user interface. Other components of the network device 20 are omitted in order not to obscure the concepts presented herein.
  • the network device 20 is in an embodiment implemented in the WCD 1, which is illustrated in Fig. 6A.
  • the WCD 1 may be provided with an
  • the WCD 1 may collect usage data by deep packet inspection through an operator of the cellular system, e.g. deep packet inspection of data, voice, voice data, or SMS.
  • the network device 20 is in an embodiment implemented in the core network 3, such as in or by an SGSN (Serving GPRS (General Packet Radio Service) Support Node), a GGSN (Gateway GPRS Support Node), a Serving Gateway, or a Packet Data Network Gateway, which is illustrated in Fig. 6B.
  • the network device 20 may in other embodiments be implemented in a Business Support System (BSS) device and/or in an Operational Support System (OSS) device, typically owned by the network operator owning the core network 3.
  • BSS Business Support System
  • OSS Operational Support System
  • the instructions of the computer program 64 may comprise a further instruction to cluster 47 a plurality of users of the cellular system together, by comparing their respective user behaviour pattern.
  • the instruction to combine 44 may comprise unite the vector value of each time interval of grouped adjacent time intervals.
  • the instruction to combine 44 may comprise average the vector value of each time interval.
  • the instruction to combine may comprise re-mine the reduced number of time intervals.
  • Each value vector may include support, confidence and lift.
  • the time intervals may be made up by one or more of the following: an hour of a day, a day of a week, a month of a year, a public holiday, and a set of one or more of hour, day, and month.
  • At least one of the time intervals may comprise at least two different data of services, and the step of defining 42 may comprise defining a value vector for each data of service for each time interval of the time intervals, by association rule mining the collected data of services.
  • FIG. 4A An embodiment of a method for identifying a user behaviour pattern in a cellular system is shown in Fig. 4A.
  • the method for identifying a user behaviour pattern in a cellular system 4 the method being performed by a network device 20 of the cellular system and comprising the steps of: collecting 40 data of services a user of the cellular system has used and at what time; dividing 42 the collected data of l8 services into time intervals; defining 42 at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services; comparing 43 the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof;
  • a user behaviour pattern is thereby identified.
  • a method for identifying a user behaviour pattern in a cellular system is shown in Fig. 4B.
  • the method may further comprise the step of clustering 47 a plurality of users of the cellular system together, by comparing their respective user behaviour pattern.
  • the step of combining 44 may comprise uniting 48 the vector value of each time interval of grouped adjacent time intervals.
  • the step of combining 44 may comprise averaging the vector value of each time interval.
  • the step of combining 44 may comprise re-mining 49 the reduced number of time intervals.
  • Each value vector may include support, confidence and lift values.
  • the time intervals may be made up by one or more of the following: an hour of a day, a day of a week, a month of a year, a public holiday, and a set of one or more of hour, day, and month.
  • At least one of the time intervals may comprise at least two different data of services, and the step of defining 42 may comprise defining a value vector for each data of service for each time interval of the time intervals, by association rule mining the collected data of services.
  • Fig. 9 illustrates an exemplifying computer system for implementing aspects of the invention. The methods as described may be implemented by means of a distributed computer system 10 (distributed data processing system).
  • Hadoop software framework is utilized in an embodiment.
  • Data e.g. call data records (CDR) 16 and/ or other types of data such as e-mail traffic, Facebook data etc., here indicated at reference numeral 17, are provided to a first computer/server 11, which maybe owned by a
  • the data is provided from the first computer/server 11 to a second computer/server 12.
  • the second computer 12 belongs to the computer system 10 and distributes the data among a group of third computers/servers 13 which also belong to the computer system 10.
  • the second computer 12 is here illustrated as a single computer, but it is noted that the second computer 12 could in fact be a set of computers
  • the distribution is thus scalable and more easily adapted to large amount of data in that the third computers 13 can work in parallel.
  • One way for implementing the distribution is to utilize Hadoop map/reduce with a master computer 18 as jobtracker.
  • the master computer 18 may then belong to the computer system as well and in such an embodiment, the master computer 18 uses both the second computer 12 and the third computers 13 as slaves/tasktrackers.
  • the number of parallel third computers 13 may be any number ranging from two computers to several thousands and using Hadoop map/reduce.
  • the computer program 64, 65 for identifying a user behaviour pattern in a cellular system comprising computer program code which, when run on a network device of a cellular system, causes the network device to: collect 40 data of services a user of the cellular system has used and at what time; divide 42 the collected data of services into time intervals;
  • the computer program product 62, 63 comprises the computer program 64, 65 and the computer readable storage means on which the computer program 64, 65 is stored.
  • the computer program can cause the processor to execute a method according to embodiments described herein.
  • the computer program product may further be an optical disc, such as a CD (compact disc) or a DVD (digital versatile disc) or a Blu-Ray disc.
  • Fig. 8 is a schematic diagram showing functional blocks of the network device 20.
  • the modules maybe implemented as only software instructions such as a computer program executing in the network device or only hardware, such as application specific integrated circuits, field programmable gate arrays, discrete logical components, transceivers, etc. or as a combination thereof.
  • some of the functional blocks may be
  • modules correspond to the steps in the methods illustrated in Figs. 4A-4B, comprising a collect manager 70, a define manager 71, a combine manage 72, and a cluster manager 73.
  • a collect manager 70 a collect manager 70
  • a define manager 71 a collect manager 70
  • a combine manage 72 a cluster manager 73.
  • these modules do not have to correspond to programming modules, but can be written as instructions according to the programming language in which they would be implemented, since some programming languages do not typically contain programming modules.
  • the collect manager 70 is configured to collect data of services a user of the cellular system has used and at what time.
  • This module corresponds to the collect step 40 of Figs. 4A and 4B.
  • This module can e.g. be implemented by the processor 30 of Fig. 5, when running the computer program.
  • the define manager 71 is configured to divide the collected data of services into time intervals, and to define at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services.
  • This module corresponds to the divide step 41 and define step 42 of Figs. 4A and 4B.
  • This module can e.g. be implemented by the processor 30 of Fig. 5, when running the computer program.
  • the combine manager 72 is configured compare the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof, and to combine the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals; wherein the combine manager further is configured to compare the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof, and to repeat combining and comparing the value vector of each time interval of the reduced time intervals, until a value vector is combined into a single time interval.
  • This module corresponds to the compare step 43, the combine step 44, the compare step 46 and the determination step 46 of Figs. 4A and 4B. This module can e.g. be
  • the cluster manager 73 is configured to cluster a plurality of users of the cellular system together, by comparing their respective user behaviour pattern.
  • This module corresponds to the cluster step 47, the unite step 48 and the generate step 49 of Fig. 4B.
  • This module can e.g. be implemented by the processor 30 of Fig. 5, when running the computer program.
  • a network device arranged to identify a user behaviour pattern in a cellular system, the network device comprising: a collect manager 70 configured to collect data of services a user of the cellular system has used and at what time; a define manager 71 configured to divide the collected data of services into time intervals, and to define at least one value vector for each time interval of the time intervals, by association rule mining the collected data of services; a combine manager 72 configured compare the value vector of each of the time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof, and to combine the value vectors of each time interval of grouped adjacent time intervals, thereby defining a value vector for each time interval of a reduced number of time intervals; wherein the combine manager further is configured to compare the value vector of each time interval of the reduced number of time intervals with the value vectors of an adjacent time interval and grouping adjacent time intervals having a minimum distance between value vectors thereof, and to repeat combining and comparing the value vector of each time interval of the
  • the network device may further comprise a cluster manager 73 configured to cluster a plurality of users of the cellular system together, by comparing their respective user behaviour pattern.
  • the combine may comprise unite the vector value of each time interval of grouped adjacent time intervals.
  • the combine may comprise average the vector value of each time interval.
  • the combine may comprise re-mine the reduced number of time intervals.
  • Each value vector may include support, confidence and lift.
  • the time intervals may be made up by one or more of the following: an hour of a day, a day of a week, a month of a year, a public holiday, and a set of one or more of hour, day, and month.
  • At least one of the time intervals may comprise at least two different data of services, and the step of defining 42 may comprise defining a value vector for each data of service for each time interval of the time intervals, by association rule mining the collected data of services.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Physics & Mathematics (AREA)
  • Educational Administration (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Game Theory and Decision Science (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Telephonic Communication Services (AREA)

Abstract

L'invention concerne un procédé permettant d'identifier un modèle de comportement d'utilisateurs dans un système cellulaire (4). Le procédé est mis en œuvre par un dispositif de réseau (20) du système cellulaire et comprend une étape consistant à collecter (40) des données indiquant les services qu'un utilisateur du système cellulaire a utilisé et à quel moment. Le procédé comprend en outre les étapes suivantes : division (42) des données de services collectées en intervalles de temps ; définition (42) d'au moins un vecteur de valeurs pour chaque intervalle de temps des intervalles de temps, en soumettant les données de services collectées à une exploration des règles d'association ; et comparaison (43) du vecteur de valeurs de chaque intervalle de temps avec les vecteurs de valeurs d'un intervalle de temps adjacent, et regroupement des intervalles de temps adjacents pour lesquels il existe une distance minimum entre les vecteurs de valeurs. Le procédé comprend également les étapes suivantes : combinaison (44) des vecteurs de valeurs de chaque intervalle de temps des intervalles de temps adjacents regroupés, ce qui permet de définir un vecteur de valeurs pour chaque intervalle de temps d'un nombre réduit d'intervalles de temps ; comparaison (45) du vecteur de valeurs de chaque intervalle de temps du nombre réduit d'intervalles de temps avec les vecteurs de valeurs d'un intervalle de temps adjacent, et regroupement des intervalles de temps adjacents pour lesquels il existe une distance minimum entre les vecteurs de valeurs ; et répétition (46) des étapes de combinaison et de comparaison du vecteur de valeur de chaque intervalle de temps du nombre réduit d'intervalles de temps jusqu'à ce qu'un vecteur de valeurs soit combiné en un intervalle de temps unique. L'invention concerne également un dispositif de réseau, un programme informatique et un produit-programme informatique correspondants .
PCT/SE2014/050629 2014-05-22 2014-05-22 Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire Ceased WO2015178811A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/SE2014/050629 WO2015178811A1 (fr) 2014-05-22 2014-05-22 Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/SE2014/050629 WO2015178811A1 (fr) 2014-05-22 2014-05-22 Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire

Publications (1)

Publication Number Publication Date
WO2015178811A1 true WO2015178811A1 (fr) 2015-11-26

Family

ID=54554361

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/SE2014/050629 Ceased WO2015178811A1 (fr) 2014-05-22 2014-05-22 Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire

Country Status (1)

Country Link
WO (1) WO2015178811A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000069194A1 (fr) * 1999-05-10 2000-11-16 Nokia Networks Oy Procede et dispositif permettant de determiner des modeles operationnels d'utilisateurs d'un systeme de telecommunications
WO2010128198A1 (fr) * 2009-05-08 2010-11-11 Zokem Oy Système et procédé d'analyse de données comportementales et contextuelles
WO2011149403A1 (fr) * 2010-05-24 2011-12-01 Telefonaktiebolaget L M Ericsson (Publ) Classification d'utilisateurs de réseau sur la base d'un comportement de réseau social correspondant
EP2568733A1 (fr) * 2010-05-20 2013-03-13 ZTE Corporation Procédé et appareil de collecte de données de communication mobile
EP2608144A2 (fr) * 2011-12-21 2013-06-26 Vodafone IP Licensing limited Catégorisation d'utilisateurs de portables basée sur la statistique de position

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000069194A1 (fr) * 1999-05-10 2000-11-16 Nokia Networks Oy Procede et dispositif permettant de determiner des modeles operationnels d'utilisateurs d'un systeme de telecommunications
WO2010128198A1 (fr) * 2009-05-08 2010-11-11 Zokem Oy Système et procédé d'analyse de données comportementales et contextuelles
EP2568733A1 (fr) * 2010-05-20 2013-03-13 ZTE Corporation Procédé et appareil de collecte de données de communication mobile
WO2011149403A1 (fr) * 2010-05-24 2011-12-01 Telefonaktiebolaget L M Ericsson (Publ) Classification d'utilisateurs de réseau sur la base d'un comportement de réseau social correspondant
EP2608144A2 (fr) * 2011-12-21 2013-06-26 Vodafone IP Licensing limited Catégorisation d'utilisateurs de portables basée sur la statistique de position

Similar Documents

Publication Publication Date Title
US11405934B2 (en) Radio resource management
US9183497B2 (en) Performance-efficient system for predicting user activities based on time-related features
US9544212B2 (en) Data usage profiles for users and applications
US20240412304A1 (en) Traveler tracking system
US11985270B2 (en) System and methods for adaptive campaign management and predictive customer engagement utilizing a customer data platform
US10148819B2 (en) Team-based customer care routing
CN103138954B (zh) 一种推荐项的推送方法、系统及推荐服务器
US20180276689A1 (en) Analyzing big data to determine a data plan
WO2014159533A1 (fr) Gestion de ressources radio améliorée
Boldrini et al. Weak signals in the mobility landscape: car sharing in ten European cities
US20200258101A1 (en) Intelligent marketing using group presence
HK1219184A1 (zh) 一种消息发送方法和设备
US20140129334A1 (en) Method and system for modeling consumer activity areas based on social media and mobile data
CN107092641A (zh) 店铺营业状态的判断方法和装置、店铺搜索的方法和装置
US20170103368A1 (en) Data processor
Ghaleb et al. Driving-situation-aware adaptive broadcasting rate scheme for vehicular ad hoc network
US9589230B1 (en) Entity location traceability and prediction
WO2015178811A1 (fr) Procédé et dispositif de réseau permettant d'identifier le modèle de comportement d'utilisateurs dans un système cellulaire
EP4446960A1 (fr) Systèmes et procédés d'identification et de réapprovisionnement d'articles ciblés sur des étagères de magasins
US20240004960A1 (en) Telecommunication network feature selection for binary classification
Liu et al. Position prediction based frequency control of beacons in vehicular ad hoc networks
Nadaf et al. Data mining in telecommunication
Pagare et al. Churn prediction by finding most influential nodes in social network
Oosterlinck et al. Home location prediction with telecom data: Benchmarking heuristics with a predictive modelling approach.
CN104750464B (zh) 一种用户状态感知、管理方法及设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14892804

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14892804

Country of ref document: EP

Kind code of ref document: A1