Fig. 3: Cross-cultural regularities across countries, language families, world subregions, and world regions identified with machine learning. | Nature Communications

Fig. 3: Cross-cultural regularities across countries, language families, world subregions, and world regions identified with machine learning.

From: Spectro-temporal acoustical markers differentiate speech from song across cultures

Fig. 3: Cross-cultural regularities across countries, language families, world subregions, and world regions identified with machine learning.The alternative text for this image may have been generated using AI.

A Left Panel: Country-wise cross-validated decoding accuracy (chance level – 50%). The colored dots represent the performance accuracy for each country (sorted as a function of accuracy with a jet colormap) n = 18 independent countries. Middle Panel: Receiver operating characteristic curve (ROC) for each country (same color code as in the left panel). Black dashed line represents the chance level. Right Panel: Features weights in the MPS domain showing features with the largest influence (z-score, average of the 18 classifiers). B–D Same as (A) for language families (n = 15 independent families), world subregions (n = 14 independent subregions), and world regions (n = 6 independent regions) respectively.

Back to article page