Fast Rule Mining Over Multi-Dimensional Windows

Abstract

Association rule mining is an indispensable tool for discovering insights from large databases and data warehouses. The data in a warehouse being multi-dimensional, it is often useful to mine rules over subsets of data defined by selections over the dimensions. Such interactive rule mining over multi-dimensional query windows is difficult since rule mining is computationally expensive. Current methods using pre-computation of frequent itemsets require counting of some itemsets by revisiting the transaction database at query time, which is very expensive. We develop a method (RMW) that identifies the minimal set of itemsets to compute and store for each cell, so that rule mining over any query window may be performed without going back to the transaction database. We give formal proofs that the set of itemsets chosen by RMW is sufficient to answer any query and also prove that it is the optimal set to be computed for 1 dimensional queries. We demonstrate through an extensive empirical evaluation that RMW achieves extremely fast query response time compared to existing methods, with only moderate overhead in pre-computation and storage.

Facilitating Interactive Mining of Global and Local Association Rules

“…Two recent works on mining associations on multidimensional data [7,28] seem relevant to our work of local rule mining. However, a closer analysis reveals that they solve a different problem altogether.…”

“…In contrast, our work focuses on a relational data model, such as in [26], where the subset attributes and the items used for forming associations are both from a common pool of attribute-value pairs. No offline assumptions are made to aid the precomputation as done in [7]. Therefore, these existing multidimensional mining approaches are inapplicable to the relational data model.…”

“…Therefore, these existing multidimensional mining approaches are inapplicable to the relational data model. All existing online mining solutions [1,7,28] compute only the support of generated rules to verify if it exceeds the user defined minsupport threshold parameter. Although minconfidence is much more expensive to verify at query-time, Wu et al [29] argue that it is critical to test a null-(transaction) invariant interestingness threshold, such as minconfidence.…”

Configurable and Extensible Multi-flows for Providing Analytics as a Service on the Cloud

“…These could include instances of commercial analytics engines such as SPSS and SAS, and toolkits of academic origin such as Cluto 10 or Weka 11 . An analytics engine takes a command from the web application to run analytics tasks on data that reside in one of the many data systems, and accordingly executes the command.…”

“…Transaction metadata based fine tuning of association rule learning, as supported by the gender and age bounds attributes in our prototype, may be used in applications to get fine-grained insights into buying behavior [11].…”