Formal Languages and Algorithms for Similarity based Retrieval from Sequence Databases

TL;DR

This paper introduces a similarity-based framework for sequence retrieval using automata, temporal logic, and regular expressions, employing distance measures to evaluate how closely sequences match queries, along with efficient algorithms for retrieval.

Contribution

It proposes a novel similarity semantics for formal query languages over sequences and provides algorithms for efficient retrieval based on these measures.

Findings

Distance measures range from 0 to 1 indicating similarity levels.

Algorithms for computing similarity are efficient and practical.

The framework enables retrieval of sequences closely matching complex queries.

Abstract

The paper considers various formalisms based on Automata, Temporal Logic and Regular Expressions for specifying queries over sequences. Unlike traditional binary semantics, the paper presents a similarity based semantics for thse formalisms. More specifically, a distance measure in the range [0,1] is associated with a sequence, query pair denoting how closely the sequence satisfies the query. These measures are defined using a spectrum of normed vector distance measures. Various distance measures based on the syntax and the traditional semantics of the query are presented. Efficient algorithms for computing these distance measure are presented. These algorithms can be employed for retrieval of sequence from a database that closely satisfy a given.