Module engram::metrics

Metrics

Functions

• cosine_similarity
  Compute the cosine similarity between two vectors. Formula: dot(a, b) / (|a| * |b|)
• jaccard_index
  Compute the Jaccard index between two sets. Formula: |A ∩ B| / |A ∪ B|
• norm_l1
  Compute the L1 norm (Manhattan distance) of a vector.
• norm_l2
  Compute the L2 norm (Euclidean distance) of a vector.
• norm_linf
  Compute the L∞ (max) norm of a vector.
• pearson_correlation
  Calculate the Pearson correlation coefficient between two vectors. Formula:
• tanimoto_coefficient
  Compute the Tanimoto Coefficient for continuous vectors. A measure of similarity for real-valued vectors that generalizes the Jaccard index. Formula: dot(a, b) / (|a|^2 + |b|^2 - dot(a, b))