Machine learning framework for detecting anomalies and insights of procurement systems
Through the machine learning framework, the multi-model method and user feedback loop self-adjustment method in the procurement system, the large amount of data and maintenance difficulties faced by the rule-based method are solved, and more efficient and accurate abnormality detection is achieved.
CN120344982APending Publication Date: 2025-07-183M INNOVATIVE PROPERTIES CO
View PDF 0 Cites 0 Cited by
Patent Information
- Application Number
- CN202380085189.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-12-20
- Filing Date
- 2023-12-19
- Publication Date
- 2025-07-18
AI Technical Summary
Technical Problem
When existing rules-based methods detect abnormalities in procurement systems, they face the problems of large data volume and diversified nature leading to design troublesome, difficult maintenance and low accuracy.
Method used
Using a machine learning framework, we map to the categories of preselected data attributes through multi-model methods, generate predictions, and make self-adjustment within the user feedback loop, use supervised learning algorithms to train the model, and combine user feedback to revise the model.
Benefits of technology
Simplifies algorithm monitoring and maintenance, improves model accuracy and adaptability, and can be quickly adjusted to new insight scenarios, providing more refined anomaly detection.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN120344982A_ABST
Abstract
Systems and techniques are described for generating machine learning models for machine learning (ML) anomaly detection and procurement. The system includes: a memory including a first data set associated with a purchase; the feature extraction module is connected to the memory, and the feature extraction module is used for extracting a first purchase feature vector set from the first purchase data set; a feature vector tagger connected to the memory, the feature vector tagger associating a tag with each feature vector in the first set of feature vectors; the machine learning modeler is connected to the feature extraction module and the feature vector labeler; and a display connected to the machine learning modeler, the display being capable of displaying to a user a purchase opportunity identified in a feature vector by the model, and being capable of receiving an indication from the user when the feature vector is misclassified.
Need to check novelty before this filing date? Find Prior Art