String similarity measures and joins with synonyms

A string similarity measure quantifies the similarity between two text strings for approximate string matching or comparison. For example, the strings "Sam" and "Samuel" can be considered similar. Most existing work that computes the similarity of two strings only considers syntactic similarities, e.g., number of common words or q-grams. While these are indeed indicators of similarity, there are many important cases where syntactically different strings can represent the same real-world object. For example, "Bill" is a short form of "William". Given a collection of predefined synonyms, the purpose of the paper is to explore such existing knowledge to evaluate string similarity measures more effectively and efficiently, thereby boosting the quality of string matching.

In particular, we first present an expansion-based framework to measure string similarities efficiently while considering synonyms. Because using synonyms in similarity measures is, while expressive, computationally expensive (NP-hard), we propose an efficient algorithm, called selective-expansion, which guarantees the optimality in many real scenarios. We then study a novel indexing structure called SI-tree, which combines both signature and length filtering strategies, for efficient string similarity joins with synonyms. We develop an estimator to approximate the size of candidates to enable an online selection of signature filters to further improve the efficiency. This estimator provides strong low-error, high-confidence guarantees while requiring only logarithmic space and time costs, thus making our method attractive both in theory and in practice. Finally, the results from an empirical study of the algorithms verify the effectiveness and efficiency of our approach.