Fig. 2: Performance of the representational prediction models. | Communications Biology

Fig. 2: Performance of the representational prediction models.

From: Hierarchical linguistic predictions and cross-level information updating during narrative comprehension

Fig. 2: Performance of the representational prediction models.

a, b Cosine distances between the predicted and actual target vectors for the word (green) and sentence (orange) prediction models, compared with the cosine distances between the predicted and randomly selected vectors (gray). c, d Classification accuracies of both word (73.234 ± 4.265%, green) and sentence predictions (81.471 ± 3.478%, orange) exceed those for random data (gray). e, f Both models exhibit an incremental context effect. The cosine distances were max–min normalized, and the vertical line in (e) indicates the average sentence length in words. Shade areas represent the standard error of the mean (SEM). g Model performance on the experimental materials.

Back to article page