R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proceedings of 5th International Conference on Knowledge Discovery and Data Mining, pages 145--154, New York, NY, 1999. ACM Press.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....intended meaning is that transactions containing X tend to contain Y too. Most (though not all) sources also require X and Y to be disjoint. We follow the standard support confidence setting since optimal rules for other interest measures can be found on the optimal support confidence border [BA]; besides, the original formulation of the Coverage Inference Scheme [CS] belonged there. The support of a rule X Y is the support of the itemset X [ Y (or its scaling when all supports are considered scaled by m) the confidence of 3 the rule is sup(X [ Y ) sup(X) The data mining process is ....

R. Bayardo, R. Agrawal. Mining the most interesting rules. In: Proc. of the 5th ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, 145--154, 1999.

....association rule describes knowledge valid in the database at hand. For instance, a confidence value of 0.7, associated to the rule above, tells that 70 percent of purchases including hamburgers and fries also include a soft drink. In the literature, several indices have been proposed (see e.g. [7], where several quality criteria are proposed) Clear enough, information patterns expressed in the form of association rules and associated indices indeed represent knowledge that might be useful in several application areas, such as, market basket analysis and fraud detection, just to mention a ....

.... allowed on boolean databases, since conditions on # values are forbidden (and a condition like A c is equivalent to A = #) We analyze the computational complexity of inducing association rules by the most frequently used rule quality indices, namely, confidence, support, # gain and h laplace [7, 2]. Specifically, we shall show that, depending on the chosen index of reference, the complexity of the problem is either in P or NP complete. When databases with nulls are considered, independently of the reference index, the rule induction task is NP complete. However, we show that there are cases ....

[Article contains additional citation context not shown here]

Roberto J. Bayardo, Jr and R. Agrawal. Mining the most interesting rules. Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, August 15-18, 1999.

....large numbers of associations. This imposes a large burden on the data analyst who must determine manually which of these associations are of interest. An active area of research is the identi cation of suitable measures of interestingness that might be applied to automatically lter associations [6, 11, 12]. This is also of importance for impact rule discovery. Aumann and Lindell [3] suggested that distribution measures be used to measure interestingness for impact rules. Their examples include the deviation from that of the training set as a whole of the mean or variance of the target for the ....

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In KDD-99, pages 145-154, 1999.

....a list of items that customers purchase in a single transaction. Many extensions to the original association rule mining [2] have been proposed to deal with some of the original algorithm s drawbacks. Examples include the use of the interestingness measure to prune down the number of rules [3][11] mining the generalized association rules involving hierarchical data set [17] and generalized affinity based association rule mining [15] 16] Most of these extensions including the original algorithm make an assumption that the data set under consideration is precise or consistent and ....

R. J. Bayardo Jr. and R. Agrawal, "Mining the Most Interesting Rules," In S. Chaudhuri and D. Madigan, editors, Proc. of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM Press, pp. 145-154, 1999.

....methods, formal concept analysis, Galois lattices. 1 Introduction Association rule mining within a transaction database [2] is a classical data mining topic, whereby the most challenging problem is the detection of frequently occurring patterns in the transaction sets (frequent itemsets) [1, 4, 11]. A major difficulty with association rules is the prohibitive number of frequent itemsets (and hence association rules) that results from even a reasonably large dataset. The frequent closed itemsets (FCIs) research topic [21, 13, 14] constitutes a promising approach to the problem of reducing ....

R.J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proceedings of the 5th International Conference on Knowledge Discovery and Data Mining (KDD'99), 1999.

....in too many rules been produced from a run of the data mining algorithm. While increasing the support and confidence level reduces the rule count, the consequence is the loss of important rules that have low support in the database [16] To overcome this, additional measures has been proposed [9, 15] to prune away rules that has been objectively identified as having no contributions to insights. Effectively, this helps the analyst focus on rules that might be useful. However, this approach has its own pitfalls. First, the use of additional measures at best achieves a reduction on the number ....

R. J. Bayardo and R. Agrawal. Mining the Most Interesting Rules. In Proc. of the 5th Int. Conf. on Knowledge Discovery and Data Mining, pages 145--154, San Diego, CA, USA, Aug. 1999.

....X Y where Xand Y are subsets of R. The most popular interestingness measure for an association rule X Y is its accuracy (or confidence) which is defined as acc(X Y, d) fr(X, d) Also several other classes of patterns and measures of interestingness have been studied (see e.g. [4, 6, 9, 13, 14, 15, 27, 32, 33, 36, 37, 43, 44, 47, 48]) It is not always easy to define an interestingness measure # in such a way that there would be a threshold value # such that #(p) # for almost all interesting patterns p and for only very few uninteresting ones. One way to augment the interestingness measure is to define additional ....

R. J. Bayardo Jr. and R. Agrawal, Mining the most interesting rules, in Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM, 1999, pp. 145--154.

....in too many rules been produced from a run of the data mining algorithm. While increasing the support and confidence level reduces the rule count, the consequence is the loss of important rules that have low support in the database [15] To overcome this, additional measures has been proposed [9, 14] to prune away rules that has been objectively identified as having no contributions to insights. Effectively, this helps the analyst focus on rules that might be useful. However, this approach has its own pitfalls. First, the use of additional measures at best achieves a reduction on the number ....

R. J. Bayardo and R. Agrawal. Mining the Most Interesting Rules. In Proc. of the 5th Int. Conf. on Knowledge Discovery and Data Mining, pages 145--154, San Diego, CA, USA, Aug. 1999.

....is frequent) makes the search easier: we know when to stop. Adding, e.g. disjunctions destroys monotonicity, and thus the search is more dicult. Finding signi cant rules etc. tends to be much harder than nding frequent patterns, although even there interesting progress has been made (see, e.g. [8]) See [19] for a fascinating approach to assigning interestingness to frequent sets. As noted above, nding frequent sets (or frequently occurring patterns from some class of patterns) is a local approach to data mining: we search for simple descriptions that are true of a reasonably large ....

R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proc. of the Fifth ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pages 145-154, 1999.

....value in doing so diminishes quickly as other aspects of KDD are weakly addressed and the analyst s goal lies in a section of the problem space. We therefore agree with [16, 18] to call for a focus on other aspects of KDD to realize the goal of data mining. Some works has been observed recently [3, 9, 33] and CrystalBall s contribution in this aspect is to eliminate the need for engineering new algorithms. Given the engine and the variant specification language, new parameters can be considered and compositions developed. This in turn eliminates programming, frees the analyst from the restrictions ....

R. J. Bayardo and R. Agrawal. Mining the Most Interesting Rules. In Proc. of ACM SIGKDD, pages 145--154, San Diego, CA, USA, August 1999.

.... interestingness of a rule, and several metrics have been proposed and used as a result of this work. 124 Among objective metrics, besides confidence and support [8] there are gain [36] variance and chi squared value [59] gini [58] strength [27] conviction [20] sc and pc optimality [13], etc. Subjective metrics include unexpectedness [73, 53, 79, 60] and actionability [66, 73, 1] Any of these metrics can be used as a part of the interestingness based filter ing tool, and the validation system can support different interestingness criteria. Moreover, the domain expert can ....

....of frequent 1 itemsets. We implemented the redundant rule elimination tool described above as a part of the validation system. However, we would like to point out that this is just one 126 of many possible redundant rule elimination approaches. Other approaches, e.g. based on ideas presented in [7, 14, 13, 55], can also be used in the rule validation process. 5.4.3 Other Validation Tools Although rule grouping and filtering proved to be the most useful and frequently used validation tool as is demonstrated in Section 5.8, they can be complemented with various other validation tools. We briefly ....

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In Pro- ceedings of the Fifth A CM $IGKDD International Conference on Knowledge Discovery and Data Mining, August 1999.

....number of frequent itemsets may be infeasible high already at comparably high levels of minimum support. So situations may occur where even quite strict constraints become essential. Such a strict constraint is the exact specification of the rule head before running the algorithm as proposed in [3]. At first glance this restriction looks quite drastic but based on it Bayardo and Agrawal are able to broaden the class of rules identified by their algorithm [3] the precise interestingness measures can be adjusted during post processing instead of a priori. So to some extend they anticipate ....

....become essential. Such a strict constraint is the exact specification of the rule head before running the algorithm as proposed in [3] At first glance this restriction looks quite drastic but based on it Bayardo and Agrawal are able to broaden the class of rules identified by their algorithm [3]: the precise interestingness measures can be adjusted during post processing instead of a priori. So to some extend they anticipate our postponed constraints approach. The general idea for such dense or otherwise problematic databases is to push as few constraints as possible into the mining. We ....

R. Bayardo and R. Agrawal. Mining the most interesting rules. In Proceedings of the 5th International Conference on Knowledge Discovery and Data Mining (KDD '99), pages 145--154, San Diego, California, USA, August 1999.

....[TKR 95,BVW00] Again, those methods do not take into consideration probabilistic interactions between rules in the cover. Also, they may prune many interesting rules if they cover instances already covered by other rules. A general study of measures of rule interestingness can be found in [BA99,JS01,HH99] An overview of the interestingness of a rule with respect to a set of constraints can be found in [GHK94] In [GHK94] the authors propose the method of random worlds and prove that in many important cases it is equivalent to the principle of maximum entropy. Maximum entropy ....

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proc. of the 5th ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pages 145-154, August 1999.

....work towards answer ing the major problem of identifying what should be considered as interesting patterns mid information within vast amounts of data. Many metrics to measure interestingness in specific contexts where suggested. Notable examples m e measuring the most interesting rules [13], mid visual feedback measuring the relevance of answers to queries [17] A major issue is whether such patterns cml be defined independently of the data set domain. It is also essential to test the validity of the patterns being discovered by the proposed meth ods. Another important issue is ....

R. J. Bayre'do Jr. and R. Agrawal. Mining the most interesting rules. In Proc. of the fifth A CM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pages 145-154, 1999.

....be applied for effectively reducing the number of candidates (or can only reduce them at the expense of loosing interesting patterns) the level wise search pioneered by Apriori might in fact not be a suitable choice. In particular if the search is performed in memory, alternative search strategies [5, 16] might be considered because of their more flexible pruning strategies. It is an open research problem which strategy is more appropriate for multi relational data mining systems like Warmr. 6 Related Work The problem of detecting temporal change in frequent patterns by grouping objects over ....

R. J. Bayardo Jr. and R. Agrawal. Mining the Most Interesting Rules. In Proceedings of the 5th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 145--154, 1999.

....deviation measures, i.e. measures of distance between association rules used for pruning similar ones, are de ned using support and con dence. Item constraints [BAG99,NLHP98] are boolean expressions that allow the user to specify the form of association rules that will be selected. In [BG99], A maximal rules, that are rules for which the population of objects concerned is reduced when an item is added to the antecedent, are de ned. In [PBTL99c] the Duquenne Guigues basis for global implications [DG86,GW99] and the Luxenburger basis for partial implications [Lux91] are adapted to the ....

R. J. Bayardo, and R. Agrawal. Mining the most interesting rules. Proc. KDD Conference, 145-154, August 1999.

....of a rule, and several metrics have been proposed and used as a result of this work. Among objective metrics, besides confidence and support [AIS93] there are gain [FMMT96] variance and chisquared value [Mor98] gini [MFM 98] strength [DT93] conviction [BMUT97] sc and pcoptimality [BA99] etc. Subjective metrics include unexpectedness [ST96b, LH96, Suz97, PT98] and actionability [PSM94, ST96b, AT97] Any of these metrics can be used as a part of the interestingness based filtering operator, and the validation system can support di#erent interestingness criteria. Moreover, the ....

....redundant rule elimination operator described above as a part of the validation system. However, we would like to point out that this redundant rule elimination operator constitutes only one type of such operator, and that other types of such operators 19 based on ideas presented in [AY98, BAG99, BA99, LHM99] can also be used in the rule validation process. 4.4 Other Validation Operators Although rule grouping and filtering proved to be the most useful and frequently used validation operators as is demonstrated in Section 6, they can be complemented with various other validation operators. ....

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, August 1999.

....easily overwhelm a human analyst. To address this problem several methods were proposed. Measures of interestingness can be used to assess the interestingness of a rule and thus generate only interesting rules or prune non interesting ones. Various such interestingness measures were discussed in [BA99] LHM99] and [JS01] for other related research we refer the reader to the excellent survey of [HH99] SLRS99] presented pruning rules for removing semantically redundant association rules. PT98] and [PT00] introduced the concept of rules that are unexpected with respect to prior beliefs and ....

Roberto J. Bayardo and Rakesh Agrawal. Mining the most interesting rules. In Proceedings of the 5th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 145-154, 1999.

....frequent patterns together with their supports for a given threshold minsup. 2 The French mathematician Blaise Pascal ( ClermontFerrand 1623, y 1662 Paris) invented an early computing device. 3 A similar notion of equivalence classes was also recently proposed by R. Bayardo and R. Agrawal [6] to characterize a maximal rules (i.e. association rules with maximal antecedent) 3. PATTERN COUNTING INFERENCE In this section, we give the theoretical basis of the new Pascal algorithm. This basis provides at the same time the proof of correctness of the algorithm. In Section 4, these ....

R. Bayardo and R. Agrawal. Mining the most interesting rules. In Proc. of the 5th Int'l Conf. on Knowledge Discovery and Data Mining (KDD), pages 145 154, Aug. 1999.

....regions, for example in Fig. 15. Treating the data in regions and differently those of the regions might be beneficial, e.g. large scale training data set could be naturally broken into smaller sets or extracting rules from the areas , and might enhance their predictive power [17]. VIII. TESTING KNOWLEDGE ACCURACY Users of the knowledge extracted from a training data set are interested in the accuracy of predictions that can be made based on this knowledge. The methods used to determine classification quality of a rule set are divided in the following three categories ....

R. J. Bayardo, Jr. and R. Agrawal, "Mining the most interesting rules," in Proc. 5th Int. ACM SIGKDD Conf. Knowledge Discovery Data Mining, 1999, pp. 145--154.

....extensive work towards answering the major problem of identifying what should be considered as interesting patterns and information within vast amounts of data. Many metrics to measure interestingness in specific contexts where suggested. Notable examples are measuring the most interesting rules [16], and visual feedback measuring the relevance of answers to queries [20] A major issue is whether such patterns can be defined independently of the data set domain. It is also essential to test the validity of the patterns being discovered by the proposed methods, since in any large volume of ....

....the interrelation or correlation between these sets of attributes. The interestingness or usefulness of the rule is usually measured by some predefined metric function such as confidence and support [2] gain [9] chi squared value [4] gini [22] entropy gain [23, 22] laplace [6, 32] lift [16], interest [5] strength [8] and conviction [5] Several proposals for mining different types of rules according to different types of pre specified interest metrics have been suggested in the literature. The suggested techniques are fully automatic but need to have predefined tasks. The ground ....

R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proc. of the fifth ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pages 145--154, 1999.

....that Pascal clearly improves the efficiency of the frequent pattern extraction from correlated data and that counting inference does not induce any execution overtime when data is weakly correlated. 2 A similar notion of equivalence classes was also recently proposed by R. Bayardo and R. Agrawal [Bay99] to characterize a maximal rules (i.e. association rules with maximal antecedent) 3 The French mathematician Blaise Pascal ( Clermont Ferrand 1623, y 1662 Paris) invented an early computing device. 3 1.3 Organization of the Paper In the next section, we recall the problem of mining ....

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. Proc. ACM SIGKDD conf., pp 145-154, 1999.

....and electronic commerce. The association rule model was introduced by Agrawal, Imielinski, and Swami [5] Starting with the pioneering work in [5] the association rule problem and its variations have been studied extensively by researchers. Several variations of the association rule problem [4, 8, 10, 19] have been proposed which can provide more interesting rules than the support confidence framework. In addition, a number of methods have been discussed in the literature which extend the binary association rule problem to related scenarios such as quantitative association rules, generalized ....

R. J. Bayardo, R. Agrawal. Mining the Most Interesting Rules. ACM SIGKDD Conference Proceedings, pages 145--154, 1999.

....de ned according to their supports and con dences, is proposed. In [BAG99,NLHP98,SVA97] the use of item constraints, that are boolean expressions de ned by the user, in order to specify the form of the association rules that will be presented to the user is proposed. The approach proposed in [BG99] is to present to the user rules with maximal antecedents, called A maximal rules, that are rules for which the population of objects concerned is reduced when an item is added to the antecedent. In [PBTL99c] we adapt the Duquenne Guigues basis for global implications [DG86,GW99] and the proper ....

R. J. Bayardo, and R. Agrawal. Mining the most interesting rules. Proc. KDD Conference, pp 145-154, August 1999.

....more than one accurate and significant description of the goal cluster. As such a set of all rules is still very large and redundant, a partial order with respect to the two criteria support and confidence is defined, where only the maximal rules with respect to this partial order are considered [3]. Finding all maximal elements is quite difficult, as the search space has to be pruned without losing significant results. We decided to skip more rules to increase performance, because the goal is not to mine all maximal rules, open list = goal cluster; results = 0; start with the most ....

R. Bayardo and R. Agrawal. Mining the most interesting rules. In S. Chaudhuri and D. Madigan, editors, Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 145--154, N.Y., 1999. ACM Press.

....rule describe knowledge valid in the database at hand. For instance a confidence value of 0.7 associated to the rule above tells that 70 percent of purchases including hamburgers and fries also include a soft drink. In the literature, several index definitions have been provided (see e.g. [7], where many interestingness criteria are proposed) Clear enough, information patterns expressed in the form of association rules and associated indices indeed denote knowledge that can be useful in several application contexts, e.g. market basket analysis. In some application contexts, however, ....

....as a special case of quantitative or categorical association rule mined on a database with nulls. In this paper, we analyze the computational complexity implied by inducing association rules using four of the mostly used rule quality indices, namely, confidence, support, V gain and laplace [7, 2]. In particular, we shall show that, in the standard case, and depending on the chosen index of reference, the complexity of the problem is either P or NP complete. When databases with nulls are considered, independently of the reference index, the rule induction task is NP complete. Despite these ....

[Article contains additional citation context not shown here]

Roberto J. Bayardo, Jr and R. Agrawal. Mining the most interesting rules. Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, August 15-18,

....are discarded. Interestingness of itemsets based on a hierarchy for the items is also discussed in [16] where for a one taxonomy situation a di erent notion of lifting to parents is used. Several other measures of interestingness for the non fuzzy case not involving taxonomies are mentioned in [2, 3, 6, 10, 15] and references in these papers; for a nice overview see [9] We would like to thank Jan Niestadt, Daniel Palomo van Es and the referees for their helpful comments. 2 Fuzzy approach If one considers more general items attributes, one has to deal with non boolean values. Several approaches have ....

R.J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In S. Chaudhuri and D. Madigan, editors, Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 145-154. ACM Press, 1999.

....mining a large database, the number of patterns discovered can easily exceed the capabilities of a human user to identify interesting results. To address this problem, various techniques have been suggested to reduce and or order the patterns prior to presenting them to the user. For example, in [3], it is shown that the most interesting rules may reside along a support confidence border. A technique is described in [20] that discovers interesting rules via an interactive process that seeks to classify rules that are not interesting. In [8] a measure is described that determines the ....

R.J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proceedings of the Fifth International Conference on Knowledge Discovery and Data Mining (KDD'99), pages 145-- 154, San Diego, California, August 1999.

....there) of discovered knowledge via the explicit detection of Simpson s paradox. An approach is described in [7] that utilizes a distance metric to evaluate the importance of a rule by considering its unexpectedness in terms of other rules in its neighborhood. Another technique is described in [3] that shows the most interesting rules reside along a support confidence border. Looking at the problem from another perspective, a method is described in [22] that attempts to discover interesting rules via an interactive process that seeks to classify rules that are not interesting. And in ....

R.J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proceedings of the Fifth International Conference on Knowledge Discovery and Data Mining (KDD'99), pages 145--154, San Diego, California, August 1999.

No context found.

R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proceedings of 5th International Conference on Knowledge Discovery and Data Mining, pages 145--154, New York, NY, 1999. ACM Press.

No context found.

R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proceedings of 5th International Conference on Knowledge Discovery and Data Mining, pages 145--154, New York, NY, 1999. ACM Press.

No context found.

R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proceedings of 5th International Conference on Knowledge Discovery and Data Mining, pages 145--154, New York, NY, 1999. ACM Press.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In 5th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining, August 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In ACM SIGKDD Conf, 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proceedings of ACM KDD'1999.

No context found.

Roberto J. Bayardo and Rakesh Agrawal. Mining the most interesting rules. In Proceedings of the fifth ACM SIGKDD international conference on knowledge discovery and data mining, pages 145--154. ACM Press, 1999.

No context found.

Bayardo, R. & Agrawal, R. (1999), Mining the most interesting rules, in `Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining', ACM Press, N.Y., pp. 145--154.

No context found.

R. Bayardo and R. Agrawal. Mining the Most Interesting Rules. In Proc. of the ACM SIGKDD Conference, 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proc. of the 5th ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pages 145--154, August 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In 5th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining, August 1999.

No context found.

R. J. Bayardo and R. Agrawal. "Mining the Most Interesting Rules." In Proc. of the 5th ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pp. 145--154, August 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In 5th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining, August 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In Proc. of the 5th ACM SIGKDD Int'l Conf. on Knowledge Discovery and Data Mining, pages 145--154, August 1999.

No context found.

R. J. Bayardo and R. Agrawal. Mining the most interesting rules. In 5th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining, August 1999.

No context found.

R. Bayardo and R. Agrawal. Mining the most interesting rules. In SIGKDD. SIGKDD, 1999.

No context found.

R. Bayardo and R. Agrawal. Mining the most interesting rules. In S. Chaudhuri and D. Madigan, editors, Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 145--154, N.Y., Aug. 15--18 1999. ACM Press.

No context found.

R. J. Bayardo Jr. and R. Agrawal. Mining the most interesting rules. In Proceedings of 5th International Conference on Knowledge Discovery and Data Mining, pages 145--154, New York, NY, 1999. ACM Press.

No context found.

R. J. Bayardo, Jr. and R. Agrawal, "Mining the Most Interesting Rules," Proceedings of the KDD, San Diego, CA, August 1999, pp. 145-154.

No context found.

Bayardo, R., Agrawal, R. "Mining the most interesting rules." Proceedings of ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD99) , pp. 145-154, 1999.

No context found.

R. J. Bayardo and R. Agrawal, "Mining the Most Interesting Rules", KDD'99, San Diego, pp.145-154, 1999.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC