Mining the most interesting rules

By Roberto J. Bayardo Jr., Rakesh Agrawal. 1999

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Sequential pattern mining -- approaches and algorithms (2013, methodology): “Constraints to guide the mining process have been employed by many algorithms, not only in sequence mining but also in association mining [Fu and Han 1995; Ng et al. 1998; Chakrabarti et al. 1998; Bayardo and Agrawal 1999; Pei et al. 2001; Ceglar et al. 2003].”
Depth first generation of long patterns (2000, background): “Sev eral variations of the association rule problem [4, 8, 10, 19] ha ve been proposed whic hcan pro vide more interesting rules than the support-con dence framew ork.”
Principles of Data Mining (2001, background): “Piatetsky-Shapiro (1991), Silberschatz and Tuzhilin (1996), and Bayardo and Agrawal (1999) contain general discussions on score functions for patterns and probabilistic rules.”
Efficient search for association rules (2000, methodology): “rithms, that have been presented as alternatives to Apriori [4, 3, 10], by exploring associations containing all available conditions as consequents.”
Generating non-redundant association rules (2000, background): “Other work addresses the problem of mining interesti ng association rules [8, 3, 10, 12].”
Mining the optimal class association rule set (2002, background): “† Too many class association rules will reduce the overall efficiency of mining predictive association rule set.”
HARMONY: Efficiently Mining the Best Rules for Classification (2005, background): “Since the introduction of association rule mining [2], many association-based (or related) classifiers have been proposed [17] [22] [7] [4] [8] [23] [15] [5] [31] [14].”
Multiobjective Optimization in Bioinformatics and Computational Biology (2007, background): “Among the most established performance measures in partial classification are the measures of coverage and confidence [27], which reflect the principles of sensitivity and specificity on a local basis.”
Alternative Interest Measures for Mining Associations in Databases (2003, background): “The conviction [8] and lift [6] metrics can be used to determine which rules, generated from a large itemset, are the most interesting.”
MAFIA: a maximal frequent itemset algorithm for transactional databases (2001, methodology): “In the second step, we use the set FI to generate “interesting” patterns, and various forms of interestingness have been proposed [8, 9, 17, 22, 27, 29, 30].”
A Survey of Temporal Knowledge Discovery Paradigms and Methods (2002, background): “For some applications, these two interestingness metrics can be misleading ([87]) and a number of other measures have been proposed [89].”