A Computational Level Theory of Similarity
Bradley C. Love · eScholarship (California Digital Library) · 2000
Why are some pairs of objects (or events) perceived to be more similar to each other than other pairs?A computational level theory of perceived similarity is presented that extends previous geometric and set-theoretic formulations.Like previous approaches, the current account posits that the similarity of two objects is a function of the common and distinctive features of the two objects.Unlike previous approaches, similarity is also a function of higher-order compatibility relations among features (as it is in models of analogy).Objects (or concepts) are represented as directed feature graphs as opposed to feature vectors or sets.Like current accounts of human analogical processing, the approach presented here holds that representational elements are put into correspondence during the comparison processes.Correspondences are chosen in order to maximize an objective function.The function contains four terms that are motivated by theories of human comparison.The maximum of the function is monotonically related to perceived similarity.Thus, similarity is characterized as the byproduct of comparison and structural alignment.The objective function serves as a quantitative computational level theory of human comparison.