Panelist Replication via Integer Least Squares for Audience Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audience measurement methods face challenges in accurately determining media audience demographics and sizes due to the lack of respondent-level data in return path data, which limits the value of media exposure information for advertisers and media providers.
Innovation Solution
The method generates synthetic respondent-level data by replicating panelists using a local minimum solution of an integer least squares problem, allowing for the creation of a seed panel that represents the entire market, even in areas without direct panel data, thereby enhancing the value of return path data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If panelists are replicated using traditional optimization procedures, then measurement precision is improved, but device complexity and processor resources increase
Solution Approach 1:
The patent changes the mathematical approach from continuous optimization to integer least squares, fundamentally altering the parameter space and solution method. This enables panelist replication through a more efficient computational framework that reduces complexity while maintaining or improving measurement precision for audience demographics
Solution Approach 2:
The patent creates synthetic panelists by copying and replicating actual panelist data according to integer weights derived from the least squares solution. This copying approach allows the system to expand limited panel data into a larger representative sample without requiring additional complex optimization procedures
2Measurement precision
If panelists are replicated using extensive processor resources, then measurement precision is improved, but productivity decreases
Solution Approach 1:
The patent transforms the computational problem into an integer least squares formulation that can be solved more efficiently than traditional optimization methods. This parameter transformation enables faster convergence to accurate panelist replication weights, improving both precision and processing speed simultaneously
3Device complexity
If local minimum solution is used instead of global optimization, then device complexity is reduced, but measurement precision may be compromised
Solution Approach 1:
The patent changes the mathematical formulation to integer least squares, which has favorable properties that allow efficient local search algorithms to find globally optimal or near-optimal solutions. This parameter transformation ensures that reduced computational complexity does not sacrifice measurement precision for demographic estimates
Data Source
AI summary
Example methods and apparatus to replicate panelists using a local minimum solution of an integer least squares problem are disclosed. An example apparatus includes memory; and processor circuitry to execute computer readable instructions to: determine weight adjustments based on a ratio of (A) a first dot product of (i) attribute data and (ii) a difference between first aggregate panelist data and second seed panelist data, and (B) a second dot product of the attribute data and the attribute data, the first aggregate panelist data based on the attribute data, the attribute data corresponding to a seed panel; determine weight estimates based on the weight adjustments, the weight estimates to replicate respective seed panelists in the seed panel; and replicate respective ones of the seed panelists based on the weight estimates to generate a synthetic audience representative of return path data reported by a plurality of media devices.