is an extension to star coordinate of Kandogan to the spherical view and for semi supervised cases i.e when a part of the data has already been annotated.
is an original approach to visualize and manipulate multidimensional data in a 2D or in 3D space. Beyond the conventional manual settings, it enriches visual clustering with automatic settings where the projection distance metric is learned from the available set of user feedback in the form of either item similarities or direct item labels.
It combines the advanced data analysis of automatic clustering with the flexibility of interactive visual clustering.  

It takes advantage of annotated items to optimize features axes.

The

Visalix

help for the Star coordinate view.

is currently used in machine learning tasks when the work of labeling data is too costly. This approach has attracted special interest in the last years.
Different methods have been developed where the user is guided during an annotation process. We propose a where the machine presents unlabeled data in an uncertainty space and lets the annotator choose the next element to label. The Model Uncertainty Visualization is both simple and powerful to visualize the nature of uncertainty for the current classifier.
This novel approach is an alternative to traditional active learning whereby the user can control the process and bring domain knowledge, insight or feedback.

The

Visalix

help for the Uncertainty view.