Reducing Computational Complexity During User Data Analysis

• US 20120214440A1
• Filed: 11/03/2009
• Published: 08/23/2012
• Est. Priority Date: 11/03/2009
• Status: Active Grant

• Current Assignee: Telefonaktiebolaget LM Ericsson
• Original Assignee: Telefonaktiebolaget LM Ericsson
• Inventors: Zahoor, J. Mohamed
• Granted Patent: US 8,626,116 B2
• Time in Patent Office: Days
• US Class Current: 455/405
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04M 15/58 based on statistics of usage or network monitoring
• CPC Class Codes: H04W 8/18 Processing of user or subscriber data, e.g. subscribed services, user preferences or user profiles; Transfer of user or subscriber data

Abstract

A method and apparatus for handling data related to users in a telecommunication network to support analysis of said users. Plural sets of parameter values relating to the users are read once, and a first level quantization cell is determined for each set of parameter values according to a default first level quantization scheme. Then, clusters of first level quantization cells populated by users are identified and an adaptive second level quantization scheme is formed by translating the identified clusters into second level quantization cells. Representative parameter values are also assigned to each second level quantization cell, wherein the representative parameter values are used for analysing the users.

Claims

1. A method of handling data related to users in a telecommunication network to support analysis of said users, the method comprising the following steps:

reading sets of parameter values obtained from traffic data of said users, each parameter value set being valid for a particular user and comprising values of two or more user-related parameters;

determining a first level quantization cell for each read set of parameter values according to a default first level quantization scheme, thereby populating first level quantization cells with users;

identifying clusters of first level quantization cells populated by users;

forming an adaptive second level quantization scheme by translating the identified clusters into second level quantization cells; and

assigning representative parameter values to each second level quantization cell, wherein the representative parameter values are used for analysing the users.

2. A method according to claim 1, wherein said first level quantization cells with populated users are stored in a first data storage, and said formation of second level quantization cells with assigned representative parameter values are stored in a second data storage.

3. A method according to claim 1, wherein the method of reading parameter values, determining first level quantization cells, and identifying clusters, is distributed over plural processing entities operating in parallel.

4. A method according to claim 3, wherein the distribution of the method over the plural processing entities operating in parallel is determined in response to the users'"'"' MSISDN number series.

5. A method according to claim 3, wherein the distribution of the method over the plural processing entities operating in parallel is determined in response to on the current load distribution in said processing entities.

6. A method according to claim 1, wherein said clusters are identified according to a predefined clustering condition dictating when a user-populated first level quantization cell belongs to a cluster.

7. A method according to claim 6, wherein the predefined clustering condition dictates that a user-populated cell belongs to a cluster when bordering at least one user-populated neighbour cell.

8. A method according to claim 7, wherein a counter of a first level quantization cell is incremented whenever that cell is populated by a user, and the predefined clustering condition further dictates that a user-populated first level quantization cell belongs to a cluster when the number of users populating the cell reaches a preset population threshold.

9. A method according to claim 1, wherein the step of identifying clusters of first level quantization cells is triggered by at least one of:

a preset timeout period has expired, a preset amount of parameter value sets have been read, and a preset amount of first level quantization cells have been populated with users.

10. A method according to claim 1, wherein one or more of the following parameter values are assigned to each second level quantization cell:

a set of parameter mean values (X,Y) and one or more sets of parameter boundary values (X_A-X_B, Y_A-Y_B, . . . ) defining the outer limits of parameter values covered by the corresponding identified cluster.

11. A data handling unit configured to handle data related to users in a telecommunication network to support analysis of said users, the data handling unit comprising:

a parameter reading unit adapted to read sets of parameter values (P) obtained from traffic data of said users, each parameter value set being valid for a particular user and comprising values of two or more user-related parameters;

a cell populating unit adapted to populate first level quantization cells with users according to a default first level quantization scheme by determining a first level quantization cell for each read set of parameter values;

a cluster identifying unit adapted to identify clusters of first level quantization cells populated by users; and

a quantization unit adapted to form an adaptive second level quantization scheme by translating the identified clusters into second level quantization cells and assigning representative parameter values to each second level quantization cell, wherein the representative parameter values are used for analysing the users.

12. A data handling unit according to claim 11, further comprising a first data storage adapted to store said first level quantization cells with populated users, and a second data storage adapted to store the formation of second level quantization cells with assigned representative parameter values.

13. A data handling unit according to claim 11, wherein the operational functions in the parameter reading unit, the cell populating unit and the cluster identifying unit for reading parameter values, determining first level quantization cells, and identifying clusters, respectively, is distributed over plural processing entities operating in parallel.

14. A data handling unit according to claim 13, wherein distribution of the operational functional is determined in response to the users'"'"' MSISDN number series.

15. A data handling unit according to claim 13, wherein distribution of the operational functional is determined in response to the current load distribution in said processing entities.

16. A data handling unit according to claim 11, wherein the cluster identifying unit is further adapted to identify said clusters according to a predefined clustering condition dictating when a user-populated first level quantization cell belongs to a cluster.

17. A data handling unit according to claim 16, wherein the predefined clustering condition dictates that a user-populated cell belongs to a cluster when bordering at least one user-populated neighbour cell.

18. A data handling unit according to claim 17, wherein the cell populating unit is further adapted to increment a counter of a first level quantization cell whenever that cell is populated by a user, and the predefined clustering condition further dictates that a user-populated first level quantization cell belongs to a cluster when the number of users populating the cell reaches a preset population threshold.

19. A data handling unit according to claim 11, wherein the cluster identifying unit is further adapted to identify clusters of first level quantization cells when triggered by at least one of:

a preset timeout period has expired, a preset amount of parameter value sets have been read, and a preset amount of first level quantization cells have been populated with users.

20. A data handling unit according to claim 11, wherein the quantization unit is further adapted to assign one or more of the following parameter values to each second level quantization cell:

a set of parameter mean values (X,Y), and one or more sets of parameter boundary values (X_A-X_B, Y_A-Y_B, . . . ) defining the outer limits of parameter values covered by the corresponding identified cluster.