Data visualization Dimensionality reduction Machine learning
In this work, we present a novel approach for data visualiza- tion based on interactive dimensionality reduction (iDR). The main idea of the paper relies on considering for visualization the intermediate results of non-convex DR algorithms under changes on the metric of the input data space driven by the user. With an appropriate visualization interface, our approach allows the user to focus on the relationships among dynamically selected groups of variables, as well as to assess the impact of a single variable or groups of variables in the structure of the data.
ESANN 2014 proceedings, European Symposium on Artificial Neural Networks, Computational Intelligence
and Machine Learning (22. 2014. Bruges (Belgium))