Audience Size Forecasting via Time Series and Frequent Item Set Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Campaign managers face challenges in accurately determining target audience size and unduplicated audience size for digital campaigns due to the complexity and size of the attribute space, leading to inefficiencies and potential budget misallocation, as existing forecasting techniques are computationally intensive and inefficient.
Innovation Solution
The system combines a time series model and a frequent item set model derived from a historic bid request database to forecast potential audience size and unduplicated audience size, using a two-phase workflow that selects a relevant time series model based on attribute values and applies categorical constraint processing to generate a frequent item set model, enabling efficient and accurate predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing forecasting techniques are used to determine target audience size and unduplicated audience size, then measurement precision is improved, but computational resources increase significantly and efficiency decreases
Solution Approach 1:
The patent segments the complex forecasting problem into distinct components: atomic target specifications are identified and processed separately, frequent item sets are mined independently from historical data, and time series models are applied to specific segments rather than the entire dataset at once. This segmentation enables parallel processing and reduces overall computational complexity while maintaining forecast accuracy.
Solution Approach 2:
The system performs preliminary actions by pre-mining frequent item sets from historical bid request data and pre-processing the data into structured formats before actual forecasting is needed. This preliminary processing creates reusable data structures that speed up subsequent forecasting operations, reducing computational resources required during campaign execution.
2Measurement precision
If comprehensive analysis of all target segments is performed to ensure accurate audience size determination, then measurement precision is improved, but time consumption increases
Solution Approach 1:
The patent extracts only the most relevant and frequent item sets from the historical data using minimum threshold filtering. By taking out and focusing only on frequent item sets that meet the threshold criteria, the system avoids analyzing all possible target segment combinations, thereby reducing forecasting time while maintaining accuracy for the most important segments.
Solution Approach 2:
The system changes parameters by adjusting the minimum threshold for frequent item set mining and selecting appropriate time series models based on data characteristics. These parameter adjustments optimize the balance between comprehensive analysis and computational efficiency, enabling accurate forecasting within acceptable timeframes.
3Measurement precision
If detailed mining of frequent item sets from historical data is performed to improve forecast accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent implements a universal frequent item set mining framework that can handle multiple types of target specifications and audience segments through a single standardized process. The same mining algorithm and data structures are used across different forecasting scenarios, reducing the need for separate complex models for each case and simplifying the overall system architecture.
Data Source
AI summary
Forecasting a potential audience size and an unduplicated audience size for a digital campaign includes receiving an audience segment input and a time period input. The audience segment input is converted into multiple atomic target specifications. For each of the multiple atomic target specifications, a potential audience size is determined during the time period input by selecting a time series model based on a frequency of attribute values from the atomic target specification and combining the selected time series model with a frequent item set model. The potential audience size for each of atomic target specifications is aggregated over the time period input into a total potential audience size. The total potential audience size is output. The time series model and the frequent item set model are obtained using data from a historic bid request database.