Class Purity
A commonly used metric in cluster analysis that measures the degree of aggregation of samples of the same class and the degree of separation between different classes in a latent space.
Timeline (last 90 days)
The experiments used lightweight models, with a total of 126 runs, and evaluation metrics included category purity and benchmark accuracy.
On unseen words and sentences, category purity increased, and this improvement generalized to samples not seen during training.
Category purity is a commonly used metric in cluster analysis: for each sample cluster corresponding to a category in the latent space, compute the proportion of samples belonging to the majority class within the cluster, then average over all clusters.
All Facts (3)
The experiments used lightweight models, with a total of 126 runs, and evaluation metrics included category purity and benchmark accuracy.
50%UnverifiedOn unseen words and sentences, category purity increased, and this improvement generalized to samples not seen during training.
50%UnverifiedCategory purity is a commonly used metric in cluster analysis: for each sample cluster corresponding to a category in the latent space, compute the proportion of samples belonging to the majority class within the cluster, then average over all clusters.
50%