Cross-Source Device Deduplication Using Behavioral Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audience measurement systems face challenges in accurately identifying and deduplicating common devices across multiple data sources, leading to potential double counting and inconsistencies in audience measurement data due to varying data quality and collection techniques.
Innovation Solution
A method for deduplicating audience measurement data by matching media devices across different data sources based on behavioral similarity, using metrics such as station duration, time match, station path, and time distance to identify common devices, without relying on device identification information like model and serial numbers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If device identification information (model and serial numbers) is used to match devices across data sources, then device identification accuracy is improved, but data privacy protection deteriorates
Solution Approach 1:
The patent extracts and removes device identification information (model and serial numbers) from the matching process. Instead of using these identifiers, the system relies on behavioral metrics such as station duration, time match, station path, and time distance to identify common devices, thereby eliminating privacy risks while maintaining identification accuracy
Solution Approach 2:
The patent introduces behavioral metrics as intermediary elements that mediate between device identification and privacy protection. These metrics (station duration, time match, station path, time distance) serve as indirect indicators that allow device matching without directly exposing or transmitting sensitive device identification information
2Quantity of substance
If multiple data sources are integrated to improve audience measurement coverage, then measurement completeness is improved, but data consistency deteriorates due to varying data quality and collection techniques
Solution Approach 1:
The patent changes the parameters used for device matching from static device identification information to dynamic behavioral metrics. By using metrics like station duration, time match, station path, and time distance that can vary and adapt across different data sources, the system achieves consistent device identification despite variations in data collection techniques and quality
Solution Approach 2:
The patent creates a universal matching mechanism using behavioral metrics that can be applied across multiple different data sources regardless of their specific collection methods. The same four metrics (station duration, time match, station path, time distance) serve as a common language for identifying common devices across diverse data sources, ensuring data consistency
3Speed
If traditional device matching methods are used to identify common devices, then identification speed is improved, but identification accuracy deteriorates due to double counting
Solution Approach 1:
The patent applies partial matching by using only four specific behavioral metrics (station duration, time match, station path, time distance) out of potentially many available device characteristics. This selective approach maintains identification speed while improving accuracy by focusing on the most discriminative metrics that effectively distinguish common devices from unique devices
Data Source
AI summary
Example methods, apparatus, systems and articles of manufacture (e.g., physical storage media) to deduplicate common devices across multiple data sources are disclosed. An example apparatus includes instructions to identify a first device in a first data source and a second device in a second data source as a possible common device, calculate at least one of a station duration metric, a time match metric or a station path metric, the station duration metric, the time match metric based times of day that the first device tuned to a second set of stations and times of day that the second device tuned to the second set of stations, determine a score based on the at least one of the station duration metric, the time match metric, or the station path metric, and determine when the first device and the second device are a common device based on the score.