Graph-Based Machine Learning for Data Linkage Confidence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Retailers face inaccuracies in linking customer data, leading to incorrect associations of customer activities, such as purchases and advertisement activities, which affects effective customer tracking and communication strategies.
Innovation Solution
A computing device is configured to obtain training data, train a machine learning algorithm, and generate confidence values for data linkages by creating graph data that associates nodes, allowing for the determination of confidence levels in linked customer data, thereby improving the accuracy of customer activity tracking and communication strategies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If customer data is linked from internal and external sources to identify customer associations, then the ability to track customer activity and attribute purchases is improved, but the accuracy of linked data deteriorates due to false associations
Solution Approach 1:
The system performs preliminary actions by collecting training data with known ground truth labels before actual data linkage operations. Machine learning models are trained in advance on this labeled data to learn patterns of accurate versus inaccurate linkages, enabling the system to predict confidence levels for future linkages before they are finalized.
Solution Approach 2:
The system implements feedback mechanisms by using ground truth labels from training data to evaluate and improve the machine learning models. The models continuously learn from the difference between predicted linkage confidence and actual accuracy, adjusting their parameters to reduce errors in future customer data linkages.
2Measurement precision
If machine learning algorithms are trained on labeled training data to determine confidence levels, then the precision of data linkage measurement is improved, but the complexity of the system increases due to training data requirements
Solution Approach 1:
The system performs preliminary actions by collecting and preparing training data with ground truth labels before deploying the machine learning models for production use. This advance preparation allows the models to be pre-trained and optimized, reducing the complexity of real-time confidence level determination while maintaining high measurement precision.
Solution Approach 2:
The machine learning models perform self-service by automatically learning from training data and improving their own accuracy without requiring manual intervention for each data linkage decision. The models self-adjust their parameters based on training feedback, reducing the operational complexity of the system while maintaining high measurement precision.
Data Source
AI summary
This application relates to apparatus and methods for determining confidence levels in associated data using machine learning algorithms. In some examples, a computing device may generate training graph data where each training graph connects at least two nodes by an edge, and each node represents data. The computing device may train a machine learning algorithm based on the generated training data. The computing device may then receive linked data, which associates at least two nodes, each representing data, with each other. The computing device may generate graph data based on the linking data, to provide to the machine learning algorithm as input. The computing device may then execute the machine learning algorithm on the generated graph data to generate values for each of its edges. The values may identify, for each edge, a confidence level in the connection between the two nodes for that edge.