One of the central problems in the eld of knowledge discovery is the development of good measures of interestingness of discovered patterns. Such measures of interestingness are divided into objective measures { those that depend only on the structure of a pattern and the underlying data used in the discovery process, and the subjective measures { those that also depend on the class of users who examine the pattern. The focus of this paper is on studying subjective measures of interestingness. These measures are classi ed into actionable and unexpected, and the relationship between them is examined. The unexpected measure of interestingness is de ned in terms of the belief system that the user has. Interestingness of a pattern is expressed in terms of how it a ects the belief system. The paper also discusses how this unexpected measure of interestingness can be used in the discovery process.

No abstract

We study the problem of discovering association rules that display regular cyclic variation over time. For example, if we compute association rules over monthly sales data, we may observe seasonal variation where certain rules are true at approximately the same month each year. Similarly, association rules can also display regular hourly, daily, weekly, etc., variation that is cyclical in nature. We demonstrate that existing methods cannot be naively extended to solve this problem of cyclic association rules. We then present two new algorithms for discovering such rules. The first one, which we call the sequential algorithm, treats association rules and cycles more or less independently. By studying the interaction between association rules and time, we devise a new technique called cycle pruning, which reduces the amount of time needed to find cyclic association rules. The second algorithm, which we call the interleaved algorithm, uses cycle pruning and other optimization techniques for discovering cyclic association rules. We demonstrate the effectiveness of the interleaved algorithm through a series of experiments. These experiments show that the interleaved algorithm can yield significant performance benefits when compared to the sequential algorithm. Performance improvements range from 5% to several hundred percent.

The history of database system research in the U.S. is one of exceptional productivity and startling economic impact. Barely twenty years old as a basic science research field, database research conducted with Federal support in the nation's universities and in its industrial research laboratories has fueled an information services industry estimated at $10 billion per year in the U.S. alone. This industry has grown at an average rate of 20 percent per year since 1965 and is continuing to expand at this rate. Achievements in database research underpin fundamental advances in communications systems, transportation and logistics, financial management, knowledge-based systems, accessibility to scientific literature, and a host of other civilian and defense applications. They also serve as the foundation for considerable progress in basic science in various fields ranging from computing to biology.

No abstract

Ahmct-Knowledpc of thc up-to-dah physical topology of an IF network Is cmcisl to 8 numhcr of crltlcai network mnnagcment tasb, Including reactive and praactlvc remum management, event carrclath, and mnt-cewe nnalysls. Given lhe dynamlc nalurc of today's IP nelworh, keeping track of topolugy infarmation manunlly Ir a dauntlng (if not Impossible) tmk. Thus, effective algnrltlims for automat i d l y dlrcovcr[np physlcnl network topolngy arc necessary, Earlier work hllrl typtcally concentrated on elther (a) dlscoverlng loglcal &@, , layer-3) topology, whkh implies thot lhr: connwtlvity of 811 Inyer-2 demcnla (cog., swltclw and bridges) b Ignored, or (b) proprletary soliitlons fargctlng spcclfic product fmiilles.In thls paper, wc pment novcl algoritlims for dhcoverlng physIca1 topology in hcterngcncous (Le., miillhvcndor) IF networks. Our algorlthms rcly on slantlard SNMP MIB Information that Ls wldcly supported by modern IP network elemcnh find rqiiire no modlflcations to the operatlng ayatem sohwarc runnlng on elements or hosts. WE Iiwe Implemented the algorithms pracntcd in thio paper in the contcxt of R topology discovery tool that has hecn tpstcd an Lucent's own rescarch nctwork. The cxperlmentnl results clearly volidalc our nppronch, demonstrating tliflt our tool cmn consistently discover the accirratc physical network topology In tlme that b mnglily qiiadmtic iii the number of nctwork elements. I. INTRODUCTIONPhysical rielwork topology refers to the characterization of the physical connectivity relationships that exist among entities in a communication network. Discovering the physical layout and interconnections of network elements is a prerequisite to many critical nctwork managcmcnt tasks, including reactive nnd proactive resource management, server siting, went correlation, and root-cause analysis. For example, consider n fault monitoring and analysis application running on a ceniral IP network management platform. qpically, a single fault in the network will causc a flood of alarm signals emanating from different interrelated network elements. Knowledge of element interconnections is essential to filter aut secondary alarm signals and correlate primary alarms to pinpoint thc original source of fai1-ut% in the network [l], [2]. Furthermore, n full physical map of the network enables a proactive analysis of thc impact of link and device failures. Early identification of single points of failiirc that could disrupt a large fraction of the user community allows the network mmager i o improve the survivabiIity of the network (e.g., by ndding nlternnte routing pnthr) bcforc outages occur.at layer-2 is definitely not straightforward. 2. Thnsparency of elements across protocol luyers. The algo-ritlim should correctly establish inlerconnections between network clcments operating ai different layers of the IS0 protocol stack. This is not trivial, since layer-2 elements in switched subncts are complctely transparcot to the layer-) router@) directing traffic in and uut of the subnets. 3. Heierageneiry of nerwark eleme...

Relaxing the assumption that relations are always in First-Normal-Form (1NF) necessitates a reexamination of the fundamentals of relational database theory. In this paper we take a first step towards unifying the various theories of ¬1NF databases. We start by determining an appropriate model to couch our formalisms in. We then define an extended relational calculus as the theoretical basis for our ¬1NF database query language. We define a minimal extended relational algebra and prove its equivalence to the ¬1NF relational calculus. We define a class of ¬1NF relations with certain “good” properties and extend our algebra operators to work within this domain. We prove certain desirable equivalences that hold only if we restrict our language to this domain.

No abstract