Abstract: A method, computer system, and a computer program product for performing a total unduplicated reach and frequency (TURF) analysis, is provided. The present invention may include distributing variables into a plurality of groups and calculating, based on a TURF analysis, a reach score for each variable in the plurality of groups. The present invention may include identifying at least one group of variables including the calculated reach score above a first threshold value and displaying a set of TURF analysis results to a user. The present invention may include determining, an available computational resource value for the TURF analysis and in response to determining that a count of variables in the identified at least one group of variables is below a second threshold value, wherein the second threshold value is based on the determined available computational resource value, performing a next TURF analysis on a single group of variables.

Abstract: A method, computer system, and a computer program product for performing a total unduplicated reach and frequency (TURF) analysis, is provided. The present invention may include distributing variables into a plurality of groups and calculating, based on a TURF analysis, a reach score for each variable in the plurality of groups. The present invention may include identifying at least one group of variables including the calculated reach score above a first threshold value and displaying a set of TURF analysis results to a user. The present invention may include determining, an available computational resource value for the TURF analysis and in response to determining that a count of variables in the identified at least one group of variables is below a second threshold value, wherein the second threshold value is based on the determined available computational resource value, performing a next TURF analysis on a single group of variables.

Abstract: Embodiments of the present invention disclose a method, computer program product, and system for performing a total unduplicated reach and frequency analysis that analyzes a large set of variables practically. A plurality of variables are received. A subset of the plurality of variables is identified as variables of interest. Variables are partitioned randomly into groups. Each of the groups of variables contains fewer than a predetermined maximum number of variables for each group. The best combination of variables within each group is found through a total unduplicated reach and frequency analysis of each group. The winning variable or variables within each group are identified based on the highest reach score. The winners from each group analysis are then pooled into a group, and the process is repeated. The combination from the second stage with the best reach is then reported. Additional variable combinations ranked by reach may also be reported.

Abstract: Embodiments of the present invention disclose a method, computer program product, and system for performing a total unduplicated reach and frequency analysis that analyzes a large set of variables practically. A plurality of variables are received. A subset of the plurality of variables is identified as variables of interest. Variables are partitioned randomly into groups. Each of the groups of variables contains fewer than a predetermined maximum number of variables for each group. The best combination of variables within each group is found through a total unduplicated reach and frequency analysis of each group. The winning variable or variables within each group are identified based on the highest reach score. The winners from each group analysis are then pooled into a group, and the process is repeated. The combination from the second stage with the best reach is then reported. Additional variable combinations ranked by reach may also be reported.

Abstract: Embodiments of the present invention disclose a method, computer program product, and system for performing a total unduplicated reach and frequency analysis that analyzes a large set of variables practically. A plurality of variables are received. A subset of the plurality of variables is identified as variables of interest. Variables are partitioned randomly into groups. Each of the groups of variables contains fewer than a predetermined maximum number of variables for each group. The best combination of variables within each group is found through a total unduplicated reach and frequency analysis of each group. The winning variable or variables within each group are identified based on the highest reach score. The winners from each group analysis are then pooled into a group, and the process is repeated. The combination from the second stage with the best reach is then reported. Additional variable combinations ranked by reach may also be reported.

Abstract: Embodiments of the present invention disclose a method, computer program product, and system for performing a total unduplicated reach and frequency analysis that analyzes a large set of variables practically. A plurality of variables are received. A subset of the plurality of variables is identified as variables of interest. Variables are partitioned randomly into groups. Each of the groups of variables contains fewer than a predetermined maximum number of variables for each group. The best combination of variables within each group is found through a total unduplicated reach and frequency analysis of each group. The winning variable or variables within each group are identified based on the highest reach score. The winners from each group analysis are then pooled into a group, and the process is repeated. The combination from the second stage with the best reach is then reported. Additional variable combinations ranked by reach may also be reported.

Abstract: A computer-implemented method, a computer-implemented system, and a computer program product for data binning are disclosed. In a first aspect, the computer-implemented method comprises in response to a determination that a data value is greater than an estimated maximum data value, calculating a new bin position for the data value by adding multiples of a bin width to the estimated maximum data value until a bin where the data value belongs is found.

Abstract: A computer-implemented method, a computer-implemented system, and a computer program product for data binning are disclosed. In a first aspect, the computer-implemented method comprises in response to a determination that a data value is greater than an estimated maximum data value, calculating a new bin position for the data value by adding multiples of a bin width to the estimated maximum data value until a bin where the data value belongs is found.

Abstract: A computer-implemented method, a computer-implemented system, and a computer program product for data binning are disclosed. In a first aspect, the computer-implemented method comprises in response to a determination that a data value is greater than an estimated maximum data value, calculating a new bin position for the data value by adding multiples of a bin width to the estimated maximum data value until a bin where the data value belongs is found.

Abstract: A computer-implemented method, a computer-implemented system, and a computer program product for data binning are disclosed. In a first aspect, the computer-implemented method comprises in response to a determination that a data value is greater than an estimated maximum data value, calculating a new bin position for the data value by adding multiples of a bin width to the estimated maximum data value until a bin where the data value belongs is found.