Machine Learning Supplier Data Association via FP Growth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional enterprise procurement applications fail to leverage the abundance of supplier data from contracts and sourcing events, relying only on filter conditions like region, commodity, and department for supplier selection without mining existing contracts and events for intelligence on the best qualified suppliers.
Innovation Solution
A machine learning-enabled procurement engine applies a frequent pattern (FP) growth tree model to identify frequent itemsets and generate association rules from supplier data, enabling the identification of suppliers that co-occur with specified attributes at above-threshold frequencies, thus recommending suitable suppliers based on user inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional filter conditions (region, commodity, department) are used for supplier selection, then the procurement process is simple and easy to operate, but the system fails to leverage the abundance of supplier data from contracts and sourcing events
Solution Approach 1:
The patent replaces conventional mechanical filtering methods with a machine learning-based frequent pattern growth tree model. This automated intelligent system analyzes supplier data automatically without requiring complex manual configuration, thus reducing information loss while maintaining operational simplicity.
Solution Approach 2:
The system enables self-service by automatically mining supplier data and generating association rules without requiring manual intervention. The frequent pattern growth tree model autonomously identifies qualified suppliers based on historical contract and sourcing event data, eliminating the need for complex manual data processing.
2Measurement precision
If machine learning model (FP growth tree) is applied to mine supplier data, then data-driven supplier selection is achieved, but the processing time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-processing supplier data into a standardized format suitable for frequent pattern growth tree analysis. This includes extracting relevant attributes from contracts and sourcing events beforehand, which reduces the computational burden during actual supplier selection and minimizes processing time.
Solution Approach 2:
The system optimizes processing efficiency by adjusting parameters such as support threshold and confidence level in the frequent pattern growth tree model. These parameter changes allow the system to balance between measurement precision (supplier qualification accuracy) and processing time based on specific procurement needs.
3Productivity
If frequent pattern growth tree is used to identify frequent itemsets, then comprehensive supplier analysis is achieved, but the system complexity and implementation difficulty increase
Solution Approach 1:
The patent segments the complex frequent pattern growth tree model into manageable components: data preprocessing module, pattern mining engine, and rule generation module. This segmentation reduces implementation difficulty while maintaining comprehensive supplier analysis capabilities and high identification efficiency.
Solution Approach 2:
The system introduces an intermediary layer that translates complex frequent pattern growth tree operations into simple supplier recommendations. This intermediary processing layer handles the computational complexity internally while presenting simplified results to users, thus improving productivity without exposing system complexity.
Data Source
AI summary
A method may include receiving a user input specifying one or more attributes. In response to the receiving the user input, a machine learning model may be applied to identify, within a supplier data stored in a database, one or more frequent itemsets containing the one or more attributes. The machine learning model may be a frequent pattern (FP) growth tree generated based on the supplier data stored in the database. A supplier having the one or more attributes may be identified based on the one or more frequent itemsets. The supplier data may include numerous contracts and/or sourcing events. As such, the supplier having the one or more attributes may be identified by leveraging the supplier data associated with the contracts and/or sourcing events. Related systems and computer program products are also provided.