Individual Contribution Attribution in Complex Data Sets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods, such as the Shapley Value Theory, struggle to accurately determine individual contributions to an outcome value in large datasets with numerous attributes, as they require specific conditions and knowledge of all possible subsets of contributors, which is often not feasible in real-world applications.
Innovation Solution
A system that determines individual contributions of entities in a chronological sequence by identifying sub-sequences with known outcome values and subtracting these from the overall outcome value, using a machine learning-based approach to analyze and attribute contributions in complex data sets, such as marketing campaigns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the Shapley Value Theory is applied to determine individual contributions, then the accuracy of contribution identification is improved, but the computational complexity and feasibility deteriorate when dealing with large datasets containing thousands of attributes
Solution Approach 1:
The patent segments the set of all contributors into multiple subsets and processes each subset separately. Instead of calculating Shapley values for all thousands of contributors simultaneously (which is computationally infeasible), the system divides them into manageable groups, calculates contributions for each group independently, and aggregates the results. This segmentation makes the computation tractable while preserving the essential contribution analysis capability.
Solution Approach 2:
The patent applies partial action by calculating Shapley values for a representative sample of subsets rather than exhaustively computing all possible permutations of all contributors. The system identifies and analyzes key subsets that capture the most significant contribution patterns, achieving satisfactory accuracy without the excessive computational burden of complete enumeration.
2Measurement precision
If the Shapley Value Theory is applied to determine individual contributions, then the accuracy of contribution identification is improved, but the data requirements and knowledge of all possible subsets deteriorate, which is not feasible in real-world applications
Solution Approach 1:
The patent segments the complete set of contributors into multiple smaller subsets, allowing the system to work with partial information about each subset rather than requiring complete knowledge of all possible combinations. This segmentation enables contribution analysis to proceed with available data while systematically handling the partial information through iterative subset analysis.
Solution Approach 2:
The patent performs preliminary identification and analysis of key subsets before attempting to determine individual contributions. By pre-identifying relevant subsets and their relationships, the system prepares the necessary partial information structure in advance, making the subsequent contribution calculation feasible even when complete subset information is unavailable.
3Productivity
If traditional contribution analysis methods are used, then the computational requirements are reduced, but the ability to handle large datasets with thousands of attributes deteriorates
Solution Approach 1:
The patent implements segmentation by dividing large datasets into manageable subsets that can be processed efficiently. Each subset is analyzed independently using contribution analysis methods, and the results are aggregated to provide overall insights. This approach maintains computational efficiency while enabling the system to scale to handle datasets with thousands of attributes that would be intractable using traditional methods.
Solution Approach 2:
The patent transitions from analyzing all contributors in a single high-dimensional space to analyzing them across multiple lower-dimensional subset spaces. By organizing contributors into hierarchical or clustered subsets, the system adds a structural dimension to the analysis, enabling efficient processing of large datasets while preserving the ability to capture complex contribution patterns across the full dataset.
Data Source
AI summary
Techniques are disclosed for determining individual contributions of corresponding individual entities that cooperatively contribute to accomplishing a goal or outcome. One technique determines an individual contribution of a target entity to an overall outcome value by identifying a first sequence of entities that includes the target entity and a corresponding sequence outcome value. Other sequences and their corresponding outcome values may be identified that partially match the first sequence while excluding the target entity. The outcome values for the other sequence(s) may be removed from the outcome value of the first sequence, thereby isolating the individual contribution of the target entity.


