## K-optimal rule discovery (2005)

Venue: | Data Mining and Knowledge Discovery |

Citations: | 17 - 3 self |

### BibTeX

@ARTICLE{Webb05k-optimalrule,

author = {Geoffrey I. Webb and Songmao Zhang},

title = {K-optimal rule discovery},

journal = {Data Mining and Knowledge Discovery},

year = {2005},

pages = {39--79}

}

### OpenURL

### Abstract

Abstract. K-optimal rule discovery finds the k rules that optimize a user-specified measure of rule value with respect to a set of sample data and user-specified constraints. This approach avoids many limitations of the frequent itemset approach of association rule discovery. This paper presents a scalable algorithm applicable to a wide range of k-optimal rule discovery tasks and demonstrates its efficiency.

### Citations

2770 | Fast Algorithms for Mining Association Rules
- Agrawal, Srikant
- 1994
(Show Context)
Citation Context ...ery tasks and demonstrates its efficiency. Keywords: Exploratory Rule Discovery, Association Rules, Classification Rules, Rule Search, Search Space Pruning 1. Introduction Association rule discovery (=-=Agrawal and Srikant, 1994-=-; Agrawal et al., 1996) is an enduring and popular data mining technology. It differs from conventional machine learning techniques by finding all rules that satisfy some set of constraints, rather th... |

1193 | Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach”, Data Mining and Knowledge Discovery - Han, Pei, et al. - 2004 |

1021 | Fast effective rule induction - Cohen - 1995 |

442 | Integrating classification and association rule mining - Liu, Hsu, et al. - 1998 |

390 | Efficiently mining long patterns from databases - Bayardo - 1998 |

382 | An efficient algorithm for mining association rules in large databases
- Savasere, Omiecinski, et al.
- 1995
(Show Context)
Citation Context ...f the four issues identified above. Most research in association rule discovery has sought to improve the efficiency of the frequent itemset discovery process (Agarwal et al., 2000; Han et al., 2000; =-=Savasere et al., 1995-=-; Toivonen, 1996, for example). This has not addressed any of the above problems, except the closed itemset approaches (Pasquier et al., 1999; Pei et al., 2000; Zaki, 2000), that reduce the number of ... |

343 | Discovering frequent closed itemsets for association rules
- Pasquier, Bastide, et al.
- 1999
(Show Context)
Citation Context ...discovery process (Agarwal et al., 2000; Han et al., 2000; Savasere et al., 1995; Toivonen, 1996, for example). This has not addressed any of the above problems, except the closed itemset approaches (=-=Pasquier et al., 1999-=-; Pei et al., 2000; Zaki, 2000), that reduce the number of itemsets required, alleviating the problems of point 3, but not addressing 1, 2 or 4. Note that auxiliary constraints, such as confidence and... |

256 | CLOSET: An efficient algorithm for mining frequent closed itemsets - Pei, Han, et al. - 2000 |

196 | Generating non-redundant association rules - Zaki - 2000 |

174 | Discovery, Analysis and Presentation of Strong Rules - Piatetsky-Shapiro - 1991 |

146 | The UCI KDD Archive. http://kdd.ics.uci.edu - Hettich, Bay - 1999 |

139 | Finding interesting associations without support pruning - Cohen, Datar, et al. |

126 | R.: Mining the most interesting rules - Bayardo, Agrawal - 1999 |

124 |
Mining association between sets of items in massive database
- Agrawal, Imielinski, et al.
- 1993
(Show Context)
Citation Context ...ituted other auxiliary constraints such as lift, also defined in Section 4.1. Most association rule discovery algorithms utilize the frequent itemset strategy as exemplified by the Apriori algorithm (=-=Agrawal et al., 1993-=-). The frequent itemset strategy first discovers all frequent itemsets {I ⊆ C | support(I, D) ≥ min support}, those sets of conditions whose support exceeds a user defined threshold min support. Assoc... |

116 | Mining Association Rules with Multiple Minimum Supports
- Liu, Hsu, et al.
- 1997
(Show Context)
Citation Context ...lost. "KORD preprint".tex; 13/12/2004; 9:27; p.4K-Optimal Rule Discovery 5 An extension of the frequent itemset approach allows min support to vary depending upon the items that an itemset contains (=-=Liu et al., 1999-=-). While this introduces greater flexibility to the frequent itemset strategy, it does not resolve any of the four issues identified above. Most research in association rule discovery has sought to im... |

113 |
Search through systematic set enumeration
- Rymon
- 1992
(Show Context)
Citation Context ...hat no superset of {b} can be a solution in the search space illustrated in Figure 1. Under previous search algorithms (Clearwater and Provost, 1990; Morishita and Nakaya, 2000; Provost et al., 1999; =-=Rymon, 1992-=-; Segal and Etzioni, 1994), the search space below such a node was pruned, as illustrated in Figure 2. In this example, pruning removes one subset from the search space. This contrasts with the prunin... |

93 | KDD-Cup 2000 organizers’ report: peeling the onion - Kohavi, Brodley, et al. |

77 | OPUS: An Efficient Admissible Algorithm for Unordered Search
- Webb
- 1995
(Show Context)
Citation Context ... system then returns the k rules that optimize λ within constraints M. This extends previous techniques that have sought the single rule that optimizes a value measure for a pre-specified consequent (=-=Webb, 1995-=-; Bayardo and Agrawal, 1999). In contrast, the new algorithm finds multiple rules and allows any condition in the role of consequent. This paper provides a formal definition of the k-optimal rule disc... |

59 | Efficient Search for Association Rules - Webb |

39 | Learning decision lists using homogeneous rules
- Segal, Etzioni
- 1994
(Show Context)
Citation Context ...et of {b} can be a solution in the search space illustrated in Figure 1. Under previous search algorithms (Clearwater and Provost, 1990; Morishita and Nakaya, 2000; Provost et al., 1999; Rymon, 1992; =-=Segal and Etzioni, 1994-=-), the search space below such a node was pruned, as illustrated in Figure 2. In this example, pruning removes one subset from the search space. This contrasts with the pruning that would occur if all... |

25 | Predictive performance of weighted relative accuracy - Todorovski, Flach, et al. - 2000 |

19 | Constraint-Based Rule Mining - Bayardo, Agrawal, et al. - 1999 |

19 | Synthesis of Optimal and Quasi-Optimal Variable-Valued Logic Formulas - Michalski - 1975 |

12 | Parallel branch-and-bound graph search for correlated association rules - Morishita, Nakaya - 2000 |

7 | Rule-space search for knowledge-based discovery
- Provost, Aronis, et al.
- 1001
(Show Context)
Citation Context ... might be determined that no superset of {b} can be a solution in the search space illustrated in Figure 1. Under previous search algorithms (Clearwater and Provost, 1990; Morishita and Nakaya, 2000; =-=Provost et al., 1999-=-; Rymon, 1992; Segal and Etzioni, 1994), the search space below such a node was pruned, as illustrated in Figure 2. In this example, pruning removes one subset from the search space. This contrasts wi... |

3 | Prasad: 2000, ‘Depth First Generation of Long Patterns - Agarwal, Aggarwal, et al. |

3 | Verkamo 1996.”Fast discovery of association rules - Agrawal, Mannila, et al. |

2 | Merz: 2001, ‘UCI Repository of Machine Learning Databases’. [Machine-readable data repository - Blake, J |

1 |
KORD preprint".tex; 13/12/2004; 9:27; p.46 Rule Discovery 47
- Clearwater, Provost
- 1990
(Show Context)
Citation Context ... subsets that cannot appear in a solution. For example, it might be determined that no superset of {b} can be a solution in the search space illustrated in Figure 1. Under previous search algorithms (=-=Clearwater and Provost, 1990-=-; Morishita and Nakaya, 2000; Provost et al., 1999; Rymon, 1992; Segal and Etzioni, 1994), the search space below such a node was pruned, as illustrated in Figure 2. In this example, pruning removes o... |