Fuzzy logic. The WWW as a data warehouse

MCIS 671 – Decision Support Systems – Spring 2005Assignment 6Pedro P. CostacurtaNova Southeastearn University 
A) Further Discussion #1 (p. 288)Discuss the utility of fuzzy logic’s ability to modelinclusion in multiple sets that intuitively are mutually exclusive orcontradictory. In what situations might this prove useful?The textbook puts it in a very clear way that weusually do not have precise limits for our descriptions but rather thatwe use categories with a range of values [5]. The modeling of this reasoningfor use by computers becomes more difficult when the ranges become blurryor intersect themselves. As an example, imagine the typical situationof driving in a highway where a speed limit is posted and the standardrecommendation is to “keep with the flow”, which, as a rule, will begoing slightly over the limit. But there will be those who will be drivingat exactly the posted value and some that will be below this number,for example elderly people or heavy trucks. In order to decide whatspeed to drive, all these inputs are considered, as well as other factorssuch as traffic entering and exiting the highway and police and or radarpresence [9], and these elements evidently show some incompatibilitybut they are, nevertheless, dealt with by our reasoning ability. Fuzzylogic extends the realm of the conventional (or Boolean logic) to incorporatethe handling of partial truth, in other words, values between 0 and1 or between true and false, something that is extremely important since“human … common sense reasoning is approximate in nature” [1]. Thisis accomplished by adding a membership value to the original “set” convertingit to what is known as a “fuzzy set”. Similarly, other extensions adoptedby fuzzy logic include fuzzy regions (relation between values and fuzzysets), hedges (“modifiers applied to fuzzy regions,”), fuzzy rules andfuzzy variables, that “are modified by the fuzzy rules and contain thecomposition of all rules executed” [7].Fuzzy logic can prove itself very useful in weatherforecasting, since approximate solutions are acceptable and importantelements can be expressed as numbers. A sample of the words normallyused in this context already demonstrate that it is not restricted toabsolute values, for example: towards midnight, low 30s, light to moderatesnow and occasional rain. Also, the “characteristic ‘hedge’ words like‘increasing’ and ‘mostly’ map nicely into the architecture of fuzzysets” [4].Business valuations can also present a good opportunityfor the use of fuzzy logic to address aspects of uncertainty. An approachto risk assessment is to consider the possibility of a given resultand fuzzy math allows the concurrent assignment of “possibilities toa number of mutually exclusive outcomes” and unlike conventional statisticalit does not have to add up to 100%. For example, one belief of, forexample 90% on a valuation of 100 does not exclude a belief of, again,90% on a valuation of 90 and “beliefs about many different valuationsover an interval would be possible” [6].Finally, another interesting use of fuzzy logicis in problems of music analysis and composition. The example presentedby Peter Elsea [2], is the concept of "just below C" whichin the conventional logic would mean that A is not an alternative andthat B and B flat are, but the fuzzy logic allows a “gradation of belowC-ness that may include A in some circumstances”. 
B) Further Discussion #1 (p. 321)Viewing the WWW as a data warehouse, describe andidentify the various warehouse components and consider how an organizationmight harness the power of the Web as a useful data warehouse.A data warehouse (DW) is typically composed ofoperational data stores that are its main source of data, metadata repositories(storing the definition of the contents of the DW and the transformationsnecessary to migrate it from the sources) and the DW itself, the actualrepository that contains cleansed data, aggregated according to thedesired business need, which can in some cases be broken down in smallerData Marts, dedicated to specific analysis.In its most simple case, the World Wide Web (WWW)can be used as a DW simply by utilizing its user interface and underlyingfacilities of search and download, with added metadata and data managementtools functionality. An example of this approach was the DistributedActive Archive Center, that provides tabular data and images for environmentalresearches and implemented a customized interface on top of its databaseof metadata, which describes the data and its location, since the actualdata is in many cases too large (e.g. satellite imagery) for a databaseor simply stored off-line [11].This usage of the WWW as a DW has expanded as someDW concepts are applied to the WWW in what can be termed “Web Warehousing”.As an example, consider the enhancement of the traditional search enginesthat are limited to keywords and do not consider the links, a majorcharacteristic of the Web, into new query mechanisms such as WebSQLthat combine database query techniques and Web retrieval methods. Similarly,the adoption of XML as a Web standard has been extended as the preferreddata extraction mechanism of information from the WWW. This trend isexemplified by the MACCS project developed by Lucent Technologies, whichalso presents an elegant distinction between data warehouses and datamarts, where the first is optimized for distribution and the secondis optimized for end-user access [10].The amazing growth of the WWW has opened a seriesof interesting possibilities of Web Mining beyond those offered by thecontent that is available, such as studies on the structure of the Webitself, for example the evolution and market penetration of technologiesand products, Web-usage, standards in eCommerce transactions and geographicaldistributions and a fascinating approach is obtained when the businessneed is analyzing the Web itself, using a DW based approach to achievethis.Within this context, three projects are worth mentioning:Web Archeology by the Systems Research Center of Compaq Computer Corporation[3] that provides a suite of tools for exploring the Web; WHOWEDA (warehouseof web data) developed at the Centre for Advanced Information Systems(CAIS) consisting of a data manipulation module responsible for gatheringinformation such as the URL, title, size and date, and an informationmining module; and the Austrian On-Line Archive (AOLA) which has theobjective of archiving the Austrian web space, consisting of the .at domain but also including other domains with servers locatedin Austria and sites related to topic [8].The AOLA project relies on web crawlers that moveacross the sites following the links, to harvest bulk data and metadata(including for example IP addresses and servers OS) from what can beseen as the operational data stores, in a process that takes monthsto complete. The results are then processed by Perl scripts, “enrichedwith information obtained from WHOIS databases” and stored in the mainrepository. The database model contains references to the web hoststhe data derives from and the hosts where links points to, and the facttable connects these parts. The main tables in the database cover theDomains, IPs, Servers, OSs, WHOIS data, pages, links, forms, filetypeand run, which provides an association with each crawl.Using OLAP tools, it was possible to constructa data cube and retrieve many different observations. Among them, thedistribution of web-servers (notably Apache and IIS) across domains,not only at a macro level but also, by drilling down, in further details,such as the presence of an open source Web server (WN Web server) basicallyin use in the academic domains. Similarly, the file types over differentWeb servers were also investigated and yielded results such as “an almostexclusive presence of the png filetype on Apache Servers, whereas almost80% of all bmp files are to be found on MS IIS servers”. 
References[1] Aziz, S. (1996).You fuzzyin' with me ?. Retrieved on May14, 2005 from http://www.doc.ic.ac.uk/~nd/surprise_96/journal/vol1/sbaa/article1.html[2] Elsea, P. (1995). Fuzzy Logic and Musical Decisions. Retrievedon May 13, 2005 from http://arts.ucsc.edu/EMS/Music/research/FuzzyLogicTutor/FuzzyTut.html[3] Leung, S., Perl, S., Stata, R. & Wiener, J. (2001). Towards web-scale web archeology. Retrievedon May 14, 2005 from. http://gatekeeper.dec.com/pub/DEC/SRC/research-reports/SRC-174.pdf.[4] Maner, W. & Joyce S. (1997). Weather Lore + Fuzzy Logic = Weather ForecastsRetrieved on May 10, 2005 from http://www.cs.bgsu.edu/maner/wxsys/wxsys.htm[5] Marakas, G. (2003). Decision Support Systems in the 21st Century.Prentice Hall. Upper Saddle River, NJ. p. 261[6] McKe, T. (2005). A New Approach to Uncertainty in Business Valuations.Retrieved on May 14, 2005 from http://www.nysscpa.org/cpajournal/2004/404/essentials/p46.htm[7] MoonJihad. (2003). An Introduction to Fuzzy Logic. Retrievedon May 10, 2005 from http://moonjihad.lns.kicks-ass.net/fuzzylogicintro.html[8] Rauber, A., Witvoet, O., Aschenbrenner, A. & Bruckner, R.(2002 ?). Putting the World Wide Web into a Data Warehouse: A DWH-based approachto Web Analysis. Retrieved on May 14, 2005 from http://citeseer.ist.psu.edu/547183.html[9] Sowell, T. (?). Fuzzy logic for just plain folks. Retrieved onMay 14, 2005 from http://www.fuzzy-logic.com/Ch1.htm[10] Varde, A. (1999). Data Warehousing and data extraction on the WorldWide Web. Retrieved on May 14, 2005 from www.cs.wpi.edu/~aparna/Webtech99.pdf[11] Yow, T., Jennings, S., Smith A., Grubb, J. & Daugherty P.(1997). Data Warehousing, Metadata, and the World Wide Web. Retrievedon May 10, 2005 from http://www.osti.gov/bridge/product.biblio.jsp?osti_id=475589

refer page:-------http://www.officesoon.com/doc/135485-fuzzy-logic-the-www-as-a-data-warehouse

File Information »

File time:2007-10-20   File size:68608   File type:doc file
Download Fuzzy logic. The WWW as a data warehouse