Visualizing word inter-relations

Enter a comma-separated list of words. We will build a map of their inter-relations in the chosen model(s), and return 2-dimensional version of this map (projected from high-dimensional vector space).

(+)
You can add new groups of words with the ’(+)’ button. They will be visualized with different colors (if there is only 1 group, the colors marks parts of speech). Optimal total number of words is from 7 to 20.

Choose the model:

Examples of large semantic maps*

5K most frequent content words in the Russian National Corpus

Plot for noun relations; click for full-size image

1K most frequent nouns in the RNC+Wikipedia corpus

Plot for noun relations; click for full-size image

* t-SNE is an algorithm for dimensionality reduction and visualization of high-dimensional datasets, developed by Laurens van der Maaten and described in this paper:
L.J.P. van der Maaten and G.E. Hinton. Visualizing High-Dimensional Data Using t-SNE. Journal of Machine Learning Research 9(Nov):2579-2605, 2008

RusVectōrēs Team

University of Oslo

Creative Commons License