topic_scaled_2d: Scaled topic coordinates in 2D space

topic_scaled_2dR Documentation

Scaled topic coordinates in 2D space

Description

Use multidimensional scaling to obtain two-dimensional coordinates for each topic in a model.

Usage

topic_scaled_2d(m, n_words = 1000)

Arguments

m

mallet_model model object

n_words

Number of columns of the topic-word matrix to use in calculation. Passed on to link{topic_divergences}.

Details

The coordinates are derived by finding the Jensen-Shannon divergences between topics considered as distributions over words and then scaling this matrix to two dimensions.

Value

a matrix with 2 columns and as many rows as m.

References

Mimno, D. 2012. Computational historiography: Data mining in a century of classics journals. ACM J. Comput. Cult. Herit. 5, no. 1 (April 2012): article 3. http://doi.acm.org/10.1145/2160165.2160168.

See Also

row_dists, topic_divergences


agoldst/dfrtopics documentation built on July 15, 2022, 4:13 p.m.