Analogy-based entropic similarity between observations and/or predictions (AESOP): an entropically weighted information-theoretic measure for benchmarking computational accuracy.
Salvador Eugenio C. CaoiliPublished in: BCB (2015)
Keyphrases
- information theoretic
- information theory
- jensen shannon divergence
- mutual information
- information theoretic measures
- similarity measure
- entropy measure
- theoretic framework
- relative entropy
- kullback leibler divergence
- information bottleneck
- jensen shannon
- kl divergence
- multi modality
- distance measure
- distance function
- shannon entropy
- log likelihood
- computational learning theory
- minimum description length
- euclidean distance
- distributional clustering
- dissimilarity measure
- semantic similarity
- image analysis
- image segmentation