Abstract
This chapter presents a discussion on fuzzy querying. It deals with the whole process of fuzzy querying, from the query formulation to its evaluation. Mainly, it advocates the use of index structures in the evaluation of fuzzy queries. First, various ways of introducing flexibility in querying processes are discussed, especially the most represented in the literature, which are based on rankings of the answers or which are using user-oriented fuzzy labels in the queries. Current methods for evaluating fuzzy queries are also reviewed. Then, properties of access methods are given in the context of fuzzy querying. Last, SaintEtiQ, the method developed in our team, is briefly presented.
Key Terms in this Chapter
Indexing Technique: Design principles, algorithms, and procedures for creating and managing an index structure. It is sometimes improperly referred to as an “index.”
Extension (of a summary): The set of tuples that belongs to the summary.
SaintEtiQ: A data summarization model, proposed by G. Raschia, that produces a hierarchy of summaries given a relational table and additional metadata.
Data Summary (or “summay”): A generalized concept, result of a classification process that groups together relational tuples whose values are close.
Index Structure (or “index”): A data structure built using data from tuple values. It is mainly used to provide fast access to records compared to a sequential scan.
User Preference: A soft requirement provided by users, in addition to a query, to reflect a wish. It influences the querying process by causing some results to be favored.
Evaluation of a Query: The process conducted for obtaining the results of a given query matched against every tuple. In flexible querying, the result is a fuzzy relation, which tuples belong to with a computed fulfillment degree.
Hierarchy of Summaries: A set of summaries organized in a tree structure using a generalization relationship between nodes of the tree.
Intension (of a summary): The expression of a summary using linguistic labels defined as fuzzy sets. It embodies the characteristics of the summary’s extension.
Top-k Answer: The set of the k best answers to a query with respect to a given result tuple evaluation scheme.