一种基于多视图深度度量学习的混合属性数据转换方法
By employing a multi-view deep metric learning method, we can fully uncover the essential characteristics of mixed attribute data and transform categorical attribute data into high-quality numerical data. This solves the problem of poor transformation in existing technologies and improves the performance and reliability of data mining and machine learning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GUIZHOU MEDICAL UNIV
- Filing Date
- 2022-09-26
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot fully uncover the essential characteristics of mixed attribute data, nor can they convert categorical attribute data into high-quality numerical data, especially when dealing with complex coupling relationships.
A multi-view depth metric learning-based approach is adopted. The multi-view information extraction module obtains coupled views of attributes within, between, and between attribute pairs and classes. The depth metric module maps these views into numerical vectors, and the fusion module merges them with numerical attribute data to form high-quality numerical data.
It achieves high-quality numerical data transformation, maintains consistent data distribution, and improves the reliability and performance of subsequent data mining and machine learning. It has low dimensionality and is suitable for mixed attribute datasets in different fields.
Smart Images
Figure CN115544137B_ABST