A Probabilistic Similarity Index

  title={A Probabilistic Similarity Index},
  author={David W. Goodall},
WILLIAMS et al.1 have recently proposed that, for similarity indices used for numerical taxonomy, each attribute should be weighted according to the sum of the χ2 values obtained from 2 × 2 tables testing the association of this attribute with all others. This relates the measure of similarity to the distance separating the two individuals in a multi-dimensional space where each axis representing an attribute has a scale depending on the importance of that attribute. Though based on χ2… CONTINUE READING