Release 1.2.0
Major Features and Improvements
New components:
- Heterogeneous feature correlation component: visualizing the correlation matrix diagram,filtering by role or variable, and scaling of the diagram
- Upload component: uploading data, and supporting for viewing data output
- Download component: downloading data, and supporting for viewing data output
- LocalBaseline component: A LR component based on Sklearn run to compare the performance of Federated LR
Component optimization:
-
Secureboost: Visual enhancement of secureboost tree diagrams, and support for scaling and dragging trees
-
Federated Sample: Filtering by label to display the table content
-
Model output of some components supports global fuzzy search
-
Model output of some components supports for table column sorting
-
Model output of some components supports multi-host visualization and filtering by role, and adds variable mapping
-
Optimize cross-validation curves and loss curves support visualization and interaction of large data volumes, and support for manual refresh during operation
Others:
- Job detail page redesign, and support for workflow scaling and dragging, and optimize parameter display
- Add job note and support for description and classification
- Log load optimization