DESC:FIELD OF THE INVENTION:
This invention relates to the field of communications engineering.

Particularly, this invention relates to a cluster formation system and method.

BACKGROUND OF THE INVENTION:
Meters are devices which are installed in a home, communicably coupled to a television, in order to capture viewing data to analyse viewership parameters.

Selection of home, and corresponding meters, has never been an easy task as it involves representation of universe (of meters and viewership data) basis primary as well as secondary control variables. This helps avoid skew at recruitment level and hence in viewership data.

Clustering of these meters is important so as to obtain pertinent demographic groups’ viewership data based on criteria set out by a user.

It is also important to determine who / which panel (meter device) can represent a current pre-defined universe (of panels). A current pre-defined universe is defined by means of parameters such as:
- State
- Town class
- Town name per town class
- NCCS
- Pin Code
- MOSR
- Mother tongue / language
- Highest Education
- Languages spoken at home
- Most often language spoken
- Household size

Therefore, there is a need for automated customised dynamic clustering of meters. This should be possible irrespective of database used for selecting clusters.

FIGURE 1 illustrates the current process for clustering being followed.

At different levels, in clustering process, different rules / filters and different systems are used. This increases manual intervention and thus increases overall time to complete work. Therefore, overall resource cost increases significantly.
While the cluster making process is mechanical work, cluster variable selection process requires intellect. The cluster making process however compromises ~80% of the time. This mechanical process requires to be automated and there is needed a system to achieve these objectives.

OBJECTS OF THE INVENTION:
An object of the invention is to define and form clusters for meter devices.

Another object of the invention is to dynamically define and form clusters for meter devices.

Yet another object of the invention is to define and form clusters for meter devices for accurate sampling.

SUMMARY OF THE INVENTION:
According to this invention, there is provided a cluster formation system, said cluster being a cluster of subscribed panels, said system comprises:
- a first definition mechanism defining a list of parameters, for panels (meter devices), through input mechanisms, for a given pre-defined geography;
- a second definition mechanism defining a list of thresholds per parameter, for panels (meter devices), through input mechanisms, for a given pre-defined geography;
- a tagger tagging said panel (meter device), based on a viewer associated with said panel, thereby defining said panel (meter device( in terms of the parameters and thresholds;
- a polling mechanism to poll each panel to determine if corresponding parameter is fully or partially available
- a filter mechanism associated with each of said parameters to filter and select a group of panels (meter devices) based on values of thresholds selected by a user through said filter mechanism, said filter mechanism being activated in order to determine clusters per defined geographic area;
- a recommendation engine configured to recommend number of possible clusters along with maximum and minimum count of homes in each cluster to enable a user to view number of clusters and number of homes in each cluster mapped to multiple filters selected.

Typically, said parameter is selected from a group of parameters consisting of a state parameter, town class parameter, town name per town class parameter, NCCS parameter, pin code parameter, MOSR parameter, mother tongue / language parameter, highest education parameter, languages spoken at home parameter, most often language spoken parameter, and household size parameter.

Typically, said cluster is a cluster of subscribed panels, said method comprising:
- identifying, by a server, requirement for clusters based on gap between currently available subscribed panels, obtained by polling said subscribed panels, and target number of panels for said cluster(s);
- determining, by a server, if parameter data, per panel, is partially available or is fully available;
- if parameter data is fully available:
o serving to a user, by a server, a variety of clusters to a user along with count in each cluster, each cluster comprising a plurality of panels (meter devices);
- if parameter data is fully available:
o computing, by a server, parameter requirements and threshold requirements in order to form a cluster;
o enabling a user, by a server, to select a variety of parameters based on said computed parameter requirements;
o enabling a user, by a server, to select thresholds per user-selected parameters based on said computed threshold requirements; and
o serving to a user, by a server, a variety of clusters to a user along with count in each cluster, each cluster comprising a plurality of panels (meter devices) selected by means of user-selected parameters and user-selected thresholds per parameter.

Typically, said parameter is selected from a group of parameters consisting of a state parameter, town class parameter, town name per town class parameter, NCCS parameter, pin code parameter, MOSR parameter, mother tongue / language parameter, highest education parameter, languages spoken at home parameter, most often language spoken parameter, and household size parameter.

Typically, information of said clusters is shared with stakeholders.

Typically, said clusters are uploaded upon approval.

Typically, said thresholds per user-selected parameters comprise stage of recruitment of a panel in a household.

Typically, said step of computing, by a server, parameter requirements and threshold requirements in order to form a cluster comprises a step of comparing currently available subscribed panels with target number of panels for said cluster(s).

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS:
FIGURE 1 illustrates the current process for clustering being followed.

The invention will now be described in relation to the accompanying drawings, in which:
FIGURE 2 illustrates a schematic block diagram of the system;
FIGURE 3 illustrates these variables and their selection;
FIGURE 4 illustrates a view of how a cluster is formed using the system; and
FIGURE 5 illustrates how a cluster is formed and uploaded using the system.

DETAILED DESCRIPTION OF THE ACCOMPANYING DRAWINGS:
According to this invention, there is provided a cluster formation system and method.

FIGURE 2 illustrates a schematic block diagram of the system of this inventon.

For an entire system, where meter devices (panels) are installed, the demographic needs to be defined.

In accordance with an embodiment of this invention, a definition mechanism defines a list of variables for a given pre-defined geography. The variables have input fields associated with them.

Afirst definition mechanism (DM1)defines a list of parameters, for panels (meter devices), through input mechanisms, for a given pre-defined geography.
Asecond definition mechanism (DM2)defines a list of thresholds per parameter, for panels (meter devices), through input mechanisms, for a given pre-defined geography.

A tagger (T) tags the input fields of the variables with a meter device, based on a viewer associated with the device, thereby defining the meter device in terms of the variables.
In at least a non-limiting embodiment, the variables / parameters may be:
- State
- Town class
- Town name per town class
- NCCS
- Pin Code
- MOSR
- Mother tongue / language
- Highest Education
- Languages spoken at home
- Most often language spoken
- Household size

Each variable must be defined as primary, secondary, or tertiary.

In at least an embodiment, a polling mechanism (PM) polls each panel (P1, P2, P3,…..,Pn) to determine if corresponding parameter is fully or partially available. Additionally, thresholds per user-selected parameters are defined, said thresholds comprise stage of recruitment of a panel in a household

FIGURE 3illustrates these variables and their selection.

A filter mechanism (FM) is associated with each of these variables to filter and select a group of meter devices based on values selected by a user through the filter mechanism.

The following steps depict how variables are selected:
- Raw data is obtained.
- Raw data is mapped with a sample design of a cluster per defined geography.
- Map the mapped sample design against an existing cluster base to check gaps in sample size per defined geography.
- Determine control variables based on this gap.
- Obtain variables from a pool of possible homes, per defined geography.
- Assign control variables to form a further cluster.
- Determine additional control variables based on data from this further cluster.
- Assign additional control variables to form an increased cluster, per geography.

Along with filter options on top, the system, of this invention, by means of a recommendation engine (RE) is configured to recommend number of possible cluster along with maximum and minimum count of homes in each cluster (basis success ratio). Once a user selects filter options, the user can see number of clusters and number of homes in each cluster mapped to multiple filters selected.

Pre-fixed Group
Variable 1/ Filter 1 Variable 2 /filter 2 Variable 3 / filter 3
XX XX XX

Number of Clusters Cluster Size Range
XX XX XX

Range Number of Cluster
XX XX
XX XX

This should be basis filters and options in those filters selected by user.

FIGURE 4 illustrates a view of how a cluster is formed using the system of this invention.

Based on control variables, a filter mechanism is activated in order to determine clusters per defined geographic area.

In at least an embodiment, the data sets / database used for cluster creation are as follows:
1. Demographic details of home including pin code must be available
2. Sample Design at State X Town X NCCS & State X Town Class
3. Control Variables proportion at State X Town Class from BI data for every variable
4. Current Sample achieved at State X Town X NCCS & State X Town Class X Control variables
5. Balance clusters / Homes for recruitment at State X Town X NCCS & State X Town Class X each of the remaining control variables
6. Proportion of control variables in already Recruited homes at State X Town Class
7. Proportion of control variable in Balance cluster homes at State X Town & Town Class level X NCCS
8. Not contacted homes for re-clustering
9. Status against all homes & individuals
10. Auto calculated data of success ratio at all primary control variables intersected. (Example : State X TC X NCCS)

The system, of this invention, accepts all kind of databases for recruitment purposes. Unused (not contacted) homes (and corresponding meters)are automatically sent for re-clustering as soon as one home from same cluster changes status to installed. The system shows count of databases available for clustering basis the sample design mapping back with UE at selected variables. The system provides auto-numbering of clusters. The system provides UE for variables and list of variables to be modified. The system provides alert for database available and sample size required for listing considering all relevant aspects are notified. All variables are part of listing data map / listing questionnaire / recruitment questionnaire. Threshold limits for primary and secondary control variables can be defined.

In at least an embodiment of this system, at the time of selection:
• At a first level, the system shows clusters based on balance sample requirement and inputs files and algorithm incorporated. This should be using primary and secondary control variables.
• Later, system considers filters as defined by research team / DQA partner and creates clusters to use. User can further add rules (threshold limits) or remove any pre-defined rules at state X Town / Town class X NCCS level.
• User has option to add /remove his / her own additional rules based on variables.
• List of variables are available in demographic data – hence user will have choice to select them along with option to save to – he / she does not have to re-run.
• Filter option for State, Region, Town Class, Town Name and NCCS is a must along with all relevant variables.
• Same home cannot be selected twice for clustering except that it is back for re-clustering:
o A home once contacted (rejected / installed / De-installed ) cannot come back for Re-clustering / clustering
o A locked Home / Address not found is not a contacted home and must come back for re-clustering / Clustering
• Pin code selection is +/- 5 including last 3 digit.
• It is possible that at the time of recruitment one of the variables for a home has changed (Example: MOSR / NCCS, etc.) Therefore, system considers the same.
• Proportion for respective variables is pulled out of recruitment file and compared with weighted proportion of the universe of homes and corresponding clusters. This is updated post every UE study.
• Clusters for casualty towns are not created and hence system does not allow the same unless the status for them changes.
• Priority is assigned to Balance Sample (Design Sample – Sample Achieved).
• Once a user has identified the combination for clusters using variables, user sees the database size and clusters getting formed to meet the sample requirement.
• Once the sample for a cell is achieved using primary control variables, system will not allow creation of additional cluster without approval.

FIGURE 5illustrates how a cluster is formed and uploaded using the system of this invention.
Step 1: The system identifies requirement for clusters based on gap between SP and SA.
Step 2: The system identifies available homes based on clusters vis-à-vis requirement. If data is partially available, move to Step 3. If data is fully available, move to Step 5. If data is not available at primary panel control level, move to Step 8.
Step 3: If data is partially available, a user checks availability of meter devices based on threshold levels set at variables.
Step 4: User sends for approval at Step 9.
Step 5: If data is available, the system shows count and combinations of clusters to a user.
Step 6: A user approves.
Step 7: Clusters get formed and uploaded. Move to Step 7 and then to Step 11.
Step 8: If data is not available at primary panel control level, the system calculates listing requirement at cell level and shares with stakeholders with an alert message.
Step 9: An approving authority approves formed clusters. Move to Step 12.
Step 10: The system generates listing requirement at cell level and shares with stakeholders.
Step 11: Information is shared with relevant stakeholders.
Step 12: Clusters get uploaded.

For the purposes of information:
• Once the number of required clusters at cell level is achieved, an automated mailer is sent highlighting the same to relevant member.
• Once approved by research they should directly flow to MDL.

For the purposes of updating:
• At any given point a home can only be used once till it is back from re-clustering
• Provision to add/ remove towns and Villages
• List of provision towns to be directly updated by MDL and intimation for the same must go to all stakeholders as defined.
• Once a Town / village has been reported as casualty town / village by MDL and approved by research than same town / village cannot be selected for making clusters as well as same must be excluded from listing requirement
• Not contacted homes by MDL must be sent for re-clustering else can be pulled by system itself.
• Multiple databases can be added and used

For the purposes of Reporting at a Macro level:
• Management summary at weekly level of sample plan vs. achieved as well as available clusters for use.
• Listing requirement that system has identified using pre-defined threshold limits at various filters

For the purposes of Reporting at a Micro level:
• At cell level on weekly basis. Including daily alerts w.r.t to required number of clusters basis sample planned Vs. Achieved.
• Alert for database shortfall with gap being less than XX% at cell level as well as state X Town Class X NCCS

The TECHNICAL ADVANCEMENT of this invention lies in providing a dynamic cluster formation of meter devices, per defined geography, in order to obtain accurate sampling. While, in the prior art, each cell (cluster) was defined as a function of state and town, in the current invention, each cell (cluster is defined) as a function of a plurality of parameters and a plurality of thresholds per parameters; thereby parameterizing panels (i.e. each panel being defined by a plurality of parameter data). Additionally, this invention makes the current system more modular since it allows addition / deletion / editing of parameters, on the fly, to form clusters; without requiring calibrations.

While this detailed description has disclosed certain specific embodiments for illustrative purposes, various modifications will be apparent to those skilled in the art which do not constitute departures from the spirit and scope of the invention as defined in the following claims, and it is to be distinctly understood that the foregoing descriptive matter is to be interpreted merely as illustrative of the invention and not as a limitation.
,CLAIMS:WE CLAIM,

1. A cluster formation system, said cluster being a cluster of subscribed panels, said system comprising:
- a first definition mechanism defining a list of parameters, for panels (meter devices), through input mechanisms, for a given pre-defined geography;
- a second definition mechanism defining a list of thresholds per parameter, for panels (meter devices), through input mechanisms, for a given pre-defined geography;
- a tagger tagging said panel (meter device), based on a viewer associated with said panel, thereby defining said panel (meter device( in terms of the parameters and thresholds;
- a polling mechanism to poll each panel to determine if corresponding parameter is fully or partially available;
- a filter mechanism associated with each of said parameters to filter and select a group of panels (meter devices) based on values of thresholds selected by a user through said filter mechanism, said filter mechanism being activated in order to determine clusters per defined geographic area; and
- a recommendation engine configured to recommend number of possible clusters along with maximum and minimum count of homes in each cluster to enable a user to view number of clusters and number of homes in each cluster mapped to multiple filters selected.

2. The cluster formation system as claimed in claim 1 wherein, said parameter being selected from a group of parameters consisting of a state parameter, town class parameter, town name per town class parameter, NCCS parameter, pin code parameter, MOSR parameter, mother tongue / language parameter, highest education parameter, languages spoken at home parameter, most often language spoken parameter, and household size parameter.

3. A cluster formation method, said cluster being a cluster of subscribed panels, said method comprising:
- identifying, by a server, requirement for clusters based on gap between currently available subscribed panels, obtained by polling said subscribed panels, and target number of panels for said cluster(s);
- determining, by a server, if parameter data, per panel, is partially available or is fully available;
- if parameter data is fully available:
o serving to a user, by a server, a variety of clusters to a user along with count in each cluster, each cluster comprising a plurality of panels (meter devices);
- if parameter data is fully available:
o computing, by a server, parameter requirements and threshold requirements in order to form a cluster;
o enabling a user, by a server, to select a variety of parameters based on said computed parameter requirements;
o enabling a user, by a server, to select thresholds per user-selected parameters based on said computed threshold requirements; and
o serving to a user, by a server, a variety of clusters to a user along with count in each cluster, each cluster comprising a plurality of panels (meter devices) selected by means of user-selected parameters and user-selected thresholds per parameter.

4. The cluster formation method as claimed in claim 1 wherein, said parameter being selected from a group of parameters consisting of a state parameter, town class parameter, town name per town class parameter, NCCS parameter, pin code parameter, MOSR parameter, mother tongue / language parameter, highest education parameter, languages spoken at home parameter, most often language spoken parameter, and household size parameter.

5. The cluster formation method as claimed in claim 1 wherein, information of said clusters are shared with stakeholders.

6. The cluster formation method as claimed in claim 1 wherein, said clusters are uploaded upon approval.

7. The cluster formation method as claimed in claim 1 wherein, said thresholds per user-selected parameters comprising stage of recruitment of a panel in a household.

8. The cluster formation method as claimed in claim 1 wherein, said step of computing, by a server, parameter requirements and threshold requirements in order to form a cluster comprising a step of comparing currently available subscribed panels with target number of panels for said cluster(s).

Dated this 08thday of January, 2020

SHILPA HEMANT GHARVE
APPLICANT’s PATENT AGENT