SHAP-Based Vendor Value Attribution for Fraud Risk Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fraud detection systems in e-commerce rely on rules-based frameworks that are inefficient in attributing value to data vendors, leading to suboptimal resource allocation and increased fraud risk.
Innovation Solution
Implementing a machine learning classification model with Shapley Additive Explanations (SHAP) to derive localized feature contributions, applying custom transformations and aggregations to determine fair value attribution for data vendors, and reallocating resources based on vendor values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If rules-based framework is used for fraud risk assessment, then implementation is simple, but fraud detection accuracy is insufficient
Solution Approach 1:
The patent replaces the mechanical rules-based framework with a machine learning classification model that uses SHAP (Shapley Additive Explanations) to evaluate vendor contributions. This substitution enables the system to achieve higher fraud detection accuracy through automated pattern recognition while maintaining interpretability through SHAP values, resolving the contradiction between implementation simplicity and detection accuracy.
2Ease of operation
If resources are allocated uniformly to data vendors, then allocation process is simple, but fraud detection effectiveness is reduced
Solution Approach 1:
The patent implements local quality by allocating resources differently to each data vendor based on their individual contribution to fraud detection. The SHAP-based evaluation mechanism determines the specific value each vendor provides, enabling differentiated resource allocation that optimizes fraud detection effectiveness while maintaining operational simplicity through automated calculation.
3Measurement precision
If SHAP-based value attribution is implemented, then vendor value assessment is accurate, but system complexity increases
Solution Approach 1:
The patent introduces SHAP (Shapley Additive Explanations) as an intermediary mechanism that bridges the gap between complex machine learning model outputs and interpretable vendor value assessments. SHAP values serve as the mediator that translates the classification model's decisions into quantifiable contributions from each vendor, achieving accurate value attribution while managing system complexity through this intermediate layer.
Data Source
AI summary
A method for managing value attribution for a data system includes obtaining, by the data system, order data associated with a plurality of data vendors, wherein the order data comprises variables about client devices, in response to obtaining the order data: applying the variables in the order data to a classification model to obtain fraud prediction models, performing an explainable mechanism on the fraud classification predictions using a shapley additive explanation (SHAP) to obtain localized SHAP values, applying a set of custom transformations to obtain normalized features, applying a value attribution algorithm on the normalized features using vendor information associated with the order data to obtain vendor values each associated with one of the plurality of data vendors, and performing a vendor remediation on the plurality of data vendors based on the obtained vendor values.