SlideShare a Scribd company logo

Dm and dw

•Download as DOCX, PDF•
•
A
Ashok kumar sandhyala

This document discusses data mining and provides information about defining data mining, explaining the knowledge discovery process with diagrams, describing a typical data mining architecture, and discussing major issues in data mining systems. It covers topics such as data cleaning, integration, selection, transformation, mining patterns, and evaluating patterns. Diagrams are provided to illustrate the knowledge discovery process and components of a data mining architecture. Major issues discussed include mining methodology, user interaction, performance, and handling diverse data types.

Related slideshows

Lecture - Data Mining
Lecture - Data MiningLecture - Data Mining
Lecture - Data Mining
 
Data mining seminar report
Data mining seminar reportData mining seminar report
Data mining seminar report
 
Data mining concepts and work
Data mining concepts and workData mining concepts and work
Data mining concepts and work
 
1 of 9
Data mining and where Housing To know technical edge with Ashok Thisis useful forall engineeringstudentshadthissubject. EspeciallyforAndhraUniversityandaffiliatedcollege IT& CSE students.Itisgood forjustone day before examand nightout students. Ashok Sandhyala 23-Jan-15
Dm and dw
Data Mining and Where Housing Assignment – 1 1. Define Data Mining?  Data Mining is extract the useful information from the large database  Data mining is extract the knowledge from the large database  Data mining is extract interesting patterns from large database  Data mining is popularly known as knowledge discovery in Database (KDD) Data miningisthe processof collecting,searchingthroughandanalyzingalarge amountof data ina database Data Miningis the practice of examininglarge pre-existingdatabasesinordertogenerate new information. 2. Explain the knowledge discovery in database with diagram?  Data miningassimplyanessential stepinte processof knowledgediscoveryindata base  Knoledge discoveryasaprocessisdepictedinfigure andcosistsof an iterative sequence of the followingsteps 1) Data cleaning:Toremove noise andinconsistentdata. 2) Data integration:Where multipledatasource maybe combined. 3) Data selection:Where datarelevanttothe analysistaskare retrevedfromthe database. 4) Data transformation:where dataare transformedorconsolidatedintoforms appropriate forminingbyperformingsummaryoraggregationoperationoperationsfor instance. 5) Data mining:Anessential processwhere intelligentmethods are appliedinorderto extractdata patterns 6) Patternevalution:Toidentifythe trulyinterestingpatterns  Knowledgetothe user  Dataminingstepmayinteractwiththe wseror knowledgebase.
 Interestingpatternsare presentedtothe use ,and maybe storage as new knowledgeinthe knowledge base.  Data miningisonlyone stepinthe entire process Since,ituncovershiddenpatternsforevolution.  Dataminingisstepinknowledgediscovery process  Data miningisbecomingmore popularthanlarge termof KDD  Data miningisthe processof discoveringinterestingknowledge.  From large amountsof data storedeitherindb’s,dw’sorotherinformationrepositories.
3. Explain a typical Data mining Architecture? : DB , DW or other informationrepository :-  Thisis one or a set of databases, dws,spreadsheetsororthe kindsof information repository.  Data cleaninganddata integrationtechniquesmaybe performedonthe data DB or DW server:-  It isresponsibleforfetchingthe relevantdatabasedonthe users data miningrequest. Knowledgebase:-  It isthe domainknowledge  Usedto guide the search, or evaluate the interestingnessof resultingpatterns.  Such knowledgecaninclude concept hierarchies, concepthierarchies,concepthierarchiesused to organize attributesorattribute valuesintodifferentlevelsof abstractions.  Knowledgesuchasuserbeliefs  Thiscan be usedto assessa patternsinterestingnessconstraintsathresholdandmetadata.
Data miningengine:-  It isessential tothe dataminingsystem  It consistsof a set of functional modules fortaskssuchas characterization,association, classificationclusteranalysis,evolutionanddeviationanalysis  Patternevaluation module:-  It employsinterestingnessmeasures  To focusinterestingpatterns  It usesinterestingpatterns  It usesinterestingnessthresholdstofilteroutdiscoveredpatterns  It integrate withminingmodule Graphical userinterface:- It communicatesbetween usersanddataminingsystem It allowingthe usertointeractwiththe systembyspecifyingadata mining It providesinformationtohelpfocusthe searchandperformingexploration dataminingbasedon the intermediatedataminingresult. From a data ware perspective:-  A data miningcanbe viewedasan advancedstage of OLAP  Datamininginvolvesanintegrationof techiniquesfronmultipledisciplinessuch 1. DB technologies 2. Statistics 3. Machine learning 4. Highperformance computing 5. Patternrecognition 6. Neutral networks 7. Data visualization 8. Informationretrieval 9. Image 10. Single processing,spatial dataanalysis  Emphasisisplacedonefficientandscalable dmtechniquesfoe large db  So dm isconsideredone of the mostimportantfrontiersindbsysteminformation industry.
4. Explain relational Data base, transactional DB, DW with diagrams? 5. Explain Advance Database systems andadvanced applications ? 6. Explain ClassificationofData Mining Systems? DM isan interdisciplinaryfield DM as a confluence of multiple disciplinesincludesDBsystem,statistics,machine learning,visualization, Informationscience. DM systemscanbe categorizedaccordingtovariouscriteria 1. Classificationaccordingtothe kindsof DB mined 2. Classificationaccordingtothe kindsof knowledgemined 3. Classificationaccordingtothe kindsof techniquesutilized 4. Classificationaccordingtothe kindsof applicationsadopted 1.Classificationaccordingtothe kindsof DB mined:-  Dm systemscanbe classified accordingtothe kindsof DB mined  DB systemsthemselvescanbe classifiedaccordingtodifferentcriterias
Suchas ---models,typesof data,applicationsinvolved.  Each criterionrequire itsownDMtechnique  DM if classifyingaccordingodatamodels.We mayhave a relationtransactional,object oriented,object-relational,Data-warehouse  If classifyingaccordingtospecial typesof datahandledwe mayhave a spatial,time ��series,text and multimedia,www 2.clssificationaccordingtothe kindsof knowledge mined:-  DM systemcategorizedaccordingtothe kindsof knowledge  CategorizedbasedonDM  DM functionalities-characterization,Discrimination,Association,classification,clustering, outlieranalysis,evolutionanalysis  A comprehensive DMsystemusuallyprovides –multiple DMfunctionalitiesintegratedDM functionalities  Move over,DMsystemscan be distinguishedbasedonthe granularityorlevel of abstractionof knowledge mined  Include-generalizedknowledge(atahigh-level of abstraction) Primitive level knowledge(ata raw data level) Knowledgeatmultiple levels(consideringseveral levelsof abstraction)  An advancedDMsystemshouldfacilitate the discoveryof knowledge atmultiple levelsof abstraction.  DM systemscanalsobe categorizedasmine dataregularitiesv/smine datairregularities. 3.classificationaccordingtothe kindsof techniquesutilized:- Categorizedaccording tothe underlyingDMtechniquesemployed Techniquescanbe describedaccordingtothe degree of userinteractive to involved Eg—autonomoussystem,interactive exploratorysystems,query-drivensystems. Technologiescanbe describedaccordingtothe methodsof data analysisemployed Eg—DB oriented , DW orientedtechniques,statistics,visualization,neutral,machine learning,pattern recognition A sophisticateddataminingsystemwill oftenadoptmultiple DMtechniquesorworkout an effective, integratedtechnique 4.classificationaccordingtothe applicationadopted:- DM systemcategorizedaccordingtothe applicationstheyadopt
Eg—DMsystemscouldbe tailoredspecificallyfoe finance,tale communication,DNA,STOCKMARKETS, E-mails. Differentapplicationsoftenrequire the integrationof applicationspecificmethods. Therefore,ageneral all-purposeDMsystemsmay not fit domain-specificminingtasks. 7. Explain Major issues in Data mining system?  Major issuesinData miningregarding 1. Miningmethodology 2. User interaction 3. Performance 4. Diverse datatypes 1.MiningmethodologyandUserinteractionissues:-  These issuesreflect  The kindsof knowledge mined  Abilitytomine knowledge atmultiple granularities  The use of domainknowledge  Ad hocDM  Knowledgevisualization a. Miningdifferentkindsof knowledgeinDM:-  Since differentuserscanbe interestedindifferentkindsof knowledge  DM shouldcovera wide spectrumof dataanalysisandknowledge discoverytasks including Data characterization,discriminatory,association,classificationtrendclustering, deviationanalysis,similarityanalysis

More Related Content

Dm and dw

  • 1. Data mining and where Housing To know technical edge with Ashok Thisis useful forall engineeringstudentshadthissubject. EspeciallyforAndhraUniversityandaffiliatedcollege IT& CSE students.Itisgood forjustone day before examand nightout students. Ashok Sandhyala 23-Jan-15
  • 3. Data Mining and Where Housing Assignment – 1 1. Define Data Mining?  Data Mining is extract the useful information from the large database  Data mining is extract the knowledge from the large database  Data mining is extract interesting patterns from large database  Data mining is popularly known as knowledge discovery in Database (KDD) Data miningisthe processof collecting,searchingthroughandanalyzingalarge amountof data ina database Data Miningis the practice of examininglarge pre-existingdatabasesinordertogenerate new information. 2. Explain the knowledge discovery in database with diagram?  Data miningassimplyanessential stepinte processof knowledgediscoveryindata base  Knoledge discoveryasaprocessisdepictedinfigure andcosistsof an iterative sequence of the followingsteps 1) Data cleaning:Toremove noise andinconsistentdata. 2) Data integration:Where multipledatasource maybe combined. 3) Data selection:Where datarelevanttothe analysistaskare retrevedfromthe database. 4) Data transformation:where dataare transformedorconsolidatedintoforms appropriate forminingbyperformingsummaryoraggregationoperationoperationsfor instance. 5) Data mining:Anessential processwhere intelligentmethods are appliedinorderto extractdata patterns 6) Patternevalution:Toidentifythe trulyinterestingpatterns  Knowledgetothe user  Dataminingstepmayinteractwiththe wseror knowledgebase.
  • 4.  Interestingpatternsare presentedtothe use ,and maybe storage as new knowledgeinthe knowledge base.  Data miningisonlyone stepinthe entire process Since,ituncovershiddenpatternsforevolution.  Dataminingisstepinknowledgediscovery process  Data miningisbecomingmore popularthanlarge termof KDD  Data miningisthe processof discoveringinterestingknowledge.  From large amountsof data storedeitherindb’s,dw’sorotherinformationrepositories.
  • 5. 3. Explain a typical Data mining Architecture? : DB , DW or other informationrepository :-  Thisis one or a set of databases, dws,spreadsheetsororthe kindsof information repository.  Data cleaninganddata integrationtechniquesmaybe performedonthe data DB or DW server:-  It isresponsibleforfetchingthe relevantdatabasedonthe users data miningrequest. Knowledgebase:-  It isthe domainknowledge  Usedto guide the search, or evaluate the interestingnessof resultingpatterns.  Such knowledgecaninclude concept hierarchies, concepthierarchies,concepthierarchiesused to organize attributesorattribute valuesintodifferentlevelsof abstractions.  Knowledgesuchasuserbeliefs  Thiscan be usedto assessa patternsinterestingnessconstraintsathresholdandmetadata.
  • 6. Data miningengine:-  It isessential tothe dataminingsystem  It consistsof a set of functional modules fortaskssuchas characterization,association, classificationclusteranalysis,evolutionanddeviationanalysis  Patternevaluation module:-  It employsinterestingnessmeasures  To focusinterestingpatterns  It usesinterestingpatterns  It usesinterestingnessthresholdstofilteroutdiscoveredpatterns  It integrate withminingmodule Graphical userinterface:- It communicatesbetween usersanddataminingsystem It allowingthe usertointeractwiththe systembyspecifyingadata mining It providesinformationtohelpfocusthe searchandperformingexploration dataminingbasedon the intermediatedataminingresult. From a data ware perspective:-  A data miningcanbe viewedasan advancedstage of OLAP  Datamininginvolvesanintegrationof techiniquesfronmultipledisciplinessuch 1. DB technologies 2. Statistics 3. Machine learning 4. Highperformance computing 5. Patternrecognition 6. Neutral networks 7. Data visualization 8. Informationretrieval 9. Image 10. Single processing,spatial dataanalysis  Emphasisisplacedonefficientandscalable dmtechniquesfoe large db  So dm isconsideredone of the mostimportantfrontiersindbsysteminformation industry.
  • 7. 4. Explain relational Data base, transactional DB, DW with diagrams? 5. Explain Advance Database systems andadvanced applications ? 6. Explain ClassificationofData Mining Systems? DM isan interdisciplinaryfield DM as a confluence of multiple disciplinesincludesDBsystem,statistics,machine learning,visualization, Informationscience. DM systemscanbe categorizedaccordingtovariouscriteria 1. Classificationaccordingtothe kindsof DB mined 2. Classificationaccordingtothe kindsof knowledgemined 3. Classificationaccordingtothe kindsof techniquesutilized 4. Classificationaccordingtothe kindsof applicationsadopted 1.Classificationaccordingtothe kindsof DB mined:-  Dm systemscanbe classified accordingtothe kindsof DB mined  DB systemsthemselvescanbe classifiedaccordingtodifferentcriterias
  • 8. Suchas ---models,typesof data,applicationsinvolved.  Each criterionrequire itsownDMtechnique  DM if classifyingaccordingodatamodels.We mayhave a relationtransactional,object oriented,object-relational,Data-warehouse  If classifyingaccordingtospecial typesof datahandledwe mayhave a spatial,time –series,text and multimedia,www 2.clssificationaccordingtothe kindsof knowledge mined:-  DM systemcategorizedaccordingtothe kindsof knowledge  CategorizedbasedonDM  DM functionalities-characterization,Discrimination,Association,classification,clustering, outlieranalysis,evolutionanalysis  A comprehensive DMsystemusuallyprovides –multiple DMfunctionalitiesintegratedDM functionalities  Move over,DMsystemscan be distinguishedbasedonthe granularityorlevel of abstractionof knowledge mined  Include-generalizedknowledge(atahigh-level of abstraction) Primitive level knowledge(ata raw data level) Knowledgeatmultiple levels(consideringseveral levelsof abstraction)  An advancedDMsystemshouldfacilitate the discoveryof knowledge atmultiple levelsof abstraction.  DM systemscanalsobe categorizedasmine dataregularitiesv/smine datairregularities. 3.classificationaccordingtothe kindsof techniquesutilized:- Categorizedaccording tothe underlyingDMtechniquesemployed Techniquescanbe describedaccordingtothe degree of userinteractive to involved Eg—autonomoussystem,interactive exploratorysystems,query-drivensystems. Technologiescanbe describedaccordingtothe methodsof data analysisemployed Eg—DB oriented , DW orientedtechniques,statistics,visualization,neutral,machine learning,pattern recognition A sophisticateddataminingsystemwill oftenadoptmultiple DMtechniquesorworkout an effective, integratedtechnique 4.classificationaccordingtothe applicationadopted:- DM systemcategorizedaccordingtothe applicationstheyadopt
  • 9. Eg—DMsystemscouldbe tailoredspecificallyfoe finance,tale communication,DNA,STOCKMARKETS, E-mails. Differentapplicationsoftenrequire the integrationof applicationspecificmethods. Therefore,ageneral all-purposeDMsystemsmay not fit domain-specificminingtasks. 7. Explain Major issues in Data mining system?  Major issuesinData miningregarding 1. Miningmethodology 2. User interaction 3. Performance 4. Diverse datatypes 1.MiningmethodologyandUserinteractionissues:-  These issuesreflect  The kindsof knowledge mined  Abilitytomine knowledge atmultiple granularities  The use of domainknowledge  Ad hocDM  Knowledgevisualization a. Miningdifferentkindsof knowledgeinDM:-  Since differentuserscanbe interestedindifferentkindsof knowledge  DM shouldcovera wide spectrumof dataanalysisandknowledge discoverytasks including Data characterization,discriminatory,association,classificationtrendclustering, deviationanalysis,similarityanalysis