Fundamentals of Similarity Search

This chapter presents various fundamental building blocks of a similarity search system from the ground up. We start by introducing the attribute-based object representation that is widely used for representing real-world entities on the web as well as in similarity search systems. We then describe the various attribute types as well as measures used to quantify similarity and dissimilarity for specific attribute types. We also touch upon the interchangeability between the notions of similarity and dissimilarity, and illustrate the various transformations that could be applied to transform a similarity measure to a measure of dissimilarity.