Evaluation of metrics(distances) between vectors.
- Euclidean distance
- Squared euclidean distance
- Cityblock distance
- Chebyshev distance
- Minkowski distance
- Hamming distance
- Cosine distance
- Correlation distance
- Spearman distance
- Kullback-Leibler divergence
- Jensen-Shannon divergence
- Many of the distances above accepts a weight vector as an optional argument to calculate weighted distances.
- The module supports computation of distances in different ways:
- compute a distance between two vectors
- compute distances between a vector and an array comprised of multiple vectors
- compute distances between corresponding vectors in two arrays along a specific dimension
- compute distances between columns in two matrices in a pairwise manner
- Specialized functions are used to compute some distances (e.g. Euclidean) in a much faster way.