Mining all frequent projection-selection queries from a relational table

Abstract

In this paper we study the problem of mining all frequent queries in a given database table, a problem known to be intractable even for conjunctive queries. We restrict our attention to projection-selection queries, and we assume that the table to be mined satisfies a set of functional dependencies. Under these assumptions we define a pre-ordering over queries and we show the following: (a) the support measure is anti-monotonic (with respect to ), and (b) if we define q ≡ q ′ iff q q ′ and q ′ q then all queries of an equivalence class have the same support.With these results at hand, we further restrict our attention to star schemas of data warehouses. In those schemas, the set of functional dependencies satisfies an important property, namely, the union of keys of all dimension tables is a key for the fact table. The main contribution of this paper is the proposal of a level-wise algorithm for mining all frequent projection-selection queries in a data warehouse over a star schema. Moreover, we show that, in the case of a star schema, the complexity in the number of scans of our algorithm is similar to that of the well known Apriori algorithm, i.e., linear with respect to the number of attributes.

Discovery and Application of Functional Dependencies in Conjunctive Query Mining

“…Recently, we have considered association rule mining on arbitrary relational databases by combining pairs of queries which could reveal interesting properties in the database [8,13]. Intuitively, we pose two queries on the database such that one query is more specific than the other (w.r.t.…”

“…Jen et al [13], studied the problem of mining all frequent queries from a single relational table. They considered projection-selection queries, and assumed that the table to be mined satisfies a set of functional dependencies.…”

“…Clearly, the combination of the approaches in [13] and [8] would resolve the issues posed, i.e., mining non redundant simple conjunctive queries (thus including arbitrary joins), given a collection of functional dependencies over the relations of an arbitrary relational database. This is one major contribution of this paper.…”

“…The paper is organised as follows: In Section 2, we recall the basic concepts and definitions used in this work and we briefly review from [13] how functional dependencies are used to compare queries. We present our algorithm Conqueror + in Section 3, combining the two approaches [8,13], and in Section 4, we report experiments, showing that Conqueror + clearly outperforms Conqueror.…”

“…We present our algorithm Conqueror + in Section 3, combining the two approaches [8,13], and in Section 4, we report experiments, showing that Conqueror + clearly outperforms Conqueror. We conclude in Section 6.…”