LW nat is KDD 8 Explain about dada, mint i in PNoLeSS © owled eo S39 Kp sta fe Knowledge discovery in Datasacae Wich f. NOLS OF ExtmacH wr Palate — Se 1S 610. OF Hue Steps inthe DD pnecess. __The folto: ving Steps ase Included in KDD process !— : = “ 4 Date cleaning 3 Date cleawing |< detined as nenaul oF noisy and Ingelevant data Gsm —Collechon . @ cleans | (ASE oF Missing Values, fi) a 4 noisy date, whene Noise S¢ a nandem on _ Vasuante esjx0n. i) cleaning. Wits Dabo diccstepancy detection and Data tyonfonmatio sooke U y Dato Tategration > Data Integsition ks detined as hetersiencout data fom multiple ___Gousrces Combined fn a Common source Data Wandios [Pata | 1 iy Misna tion Data __ a} Dake selection J Data Selection fs detined as | ___sthe process thane date velevant . do the analye?s is decided and setsdeved from be _ data collecHon. Fos. tus we can use Nousal network, Teacher's Signature.—__ pecision | Tees, Naive —___ Pegnessiom Wmethod.t . ea aeeeeeeeeeeat is defined as Hire prev, 106 +1anfooiney ata | sopKite fox Sega sie ina _by procedune. Data. Tron: - : fosmation ts a dwo Hep pTocess O Data mapping 3 Assigning elements for sour : base do dettnation to Gptune es _drantfosumatione . ———} ® Cede Yenenahon 3 Creation of toe aclual | does fosmation progoram . 2 Dato fring, Date wining is defined as- [ ni “ase i 7 l act wn potent 2 saenioiiae eee Nelova: i en eades _purpore of Yandel leg clastic oon 6) pattem Evaluation %— Pattern Evaluation is. deéined as idenbityiv g stedictly : incneati pabtesine srepsresenting Knowle dae loosed) given wmeasusrer Tt Elna {ntosecti a —_|| Crome renn, And Uses summaxisat Arnel Visuadi2ation do make data undexstandable _ by 4 FARE Repmescntation s— Thy invalves pres = ling dhe sresult In a wa thot weaning ful--and co" rong be used +o make dactink | 3 eeQusshennestes, } ylhat ts Mativated Data rnining 9 Data muning is metiyated by She following rease ns Motivation { impostance 4) Mastcket Analysis > doget a mone holistic vj ew oun Client 1: data oaburg_and mu eh ce ge swe can (wan Woe about customise dastes with data fake @ look at puachare btn and | ucla tm mone, we can then have mos ustpmized “ustomen expesdences wir Hails elning neseanch, update ous wasketing. Stnatesy setain o SUAOs10US analysis PHLOCeSS and Piel Sola fo which” Customons gore mere to 4) aud Detect > “usage OF one’s cameo for - _pensonal neasone ensuchment by the Wwoliclous misuse og execution of tne Wealth om pyro Ves OF Hne nrecgutti any " | n0l04}«. _Sastems have olishonesd processes This has happened in yiany aspect oF eve life, Such as Nebwotk ‘Tek eo) ica wobile commu Catt i = 4) customen Retention 5 “the setention of cu sotention of cus iz applies to a business 0% P “obi ily do fo maimtarn 14s Customers foo a give _pessed High ee of ee means that buyes dus o€ fire arioduct on Compare porefesr do Setusier “Continue Ih oD Shop 09 othesuwice not defect do another: prada. OF Company oa mot do use Mt cltope dient: aeTosaite spp cee SONI) [rou lo ing -and_deaniny daka. ane nea, ty “supe. wtrgouizations “have input seconds but ae eer ay aa ; it ice VcomPanres Wave a Long hice Challe: etiins up Spenating preckdunes to ath poleduchion puiececsces 5) schentitic ExPlonatten 8— Data dis coves. isa | : Iced close do tnitiat date anudycis , wh a date scientist ses Visual | ©: lGuatimn water, then cmventi{onat date prowess: os ee ee func onality of tue data uch featuses can [nduele ata Size 09 mers , data consicten. _ potential intenactions between data _olements on data: Files ( tales. ‘Discuss Ghout Data Mining Taak podwd Fives [excumple Qa a Pr. P Data tani t “AO prindtives ame the pasic — socclty hat kind et data donnie, wht! u ime_a_dlata wlwiny oe ecieare sacs hetot knowledge 40 aliscovex, what backs mound Pnowled go use Mow to beastie fue ind enestingness OF tne . a and how de present the stesults Data Mining tar prdvaitives can be used ts Consforuct a data minin ee wbich 1s a foormag. Wood OF exporessing Jue Use9's ae “A Y ino, goals . Fo 51 Cxample SUPPOSE a user wants +o “to find thea shade 1 seules, re the [tern Sold ina supemmarket 9 possible di SIT tas) irra 9 Sek oF task-selevant dake > The sales Huntachons in the supesimasicet database . whane ach tnansaction conRing a set of /tems puncha seal by QA cusdomose © Kind of Knowledge to be mined D> AssociaHen sues, with ane nutes of tre wn oy, whene Kandy ane sets of (tems , meaning that customers wh buy X alco tend te buy ¥. : © Background Knowledge > “Te vaimimum Support and Cméidence Haneshalde, _ “which ane sed do Fltese out the Suules Hat agie not iment 0% gwllahle enougls. _ _ (he, User. aay also ; species “oben coditeoda —F eyat ute she Senter, suet af e+ comeiction Jes ——_] 0%. 1 ae sant do find Lt 9, Eyaurpla EHE Usegt way wat ot novelty Fe f ty yeas that | Ss aa that have a big WE wo ai| the association between ¥ pte —____ expected hy chance 4 _@ Representation foo visualizing foe dicovensy 1 Pattes urs 3- : Trayec Sse may tease fo see Hoe 7 Wisc tn a balan. tuto ohana enc How ——shews tne antecedent, consequent, suppost, Confidence and ether meacumnes of a yule Alten. Ae seo mas for 40 SCE the HUI in a i : with @ label Showing she appa ol ___|_ Confidence } explain in detail about Date airing Eunctiona- ities 7 as =) Data wining functionalities ane Hee tyes « paltouss osc Bhowle that can be discovered Tsim date. using data imivimg bechwuques Theme ane ditFement Kinds of data Waning 7 farnchonalities }— “Classification 3 pic Zanetimality ‘s.osed t= Assign a labet 09 chitegors 4 fata trotance based on iss foadusees 7 yee b H Crapper - Classification can be| ell buy a a. erie os Nat. sre con | foo done ust een sucha a6 decision Tree) “net vx On 0 machin ie a predichen D> The Funchghabity js useo ¥ @ ia nielaHonslups fos dependencies am i yasddoles ov ifems. zt \ he funchonal I's used to estimate bie valuoree le basecl on tre values OF ofuen Vasddlle Ve shale + Serer can be useol +o forecast the Sales of a Product, tnepsice of a Stock om the yisk of a disease. prediction can be done using ‘vasous techniques » Cuch as sregeession analysis, dime Seades jomalysis , 0% K- heanesd neighbors ( | ® Pssociation analysis 2 Sh one a nelationshps on ___| dependena on le items in Jue d | Fost example» association amalysl can be used +o tl di couse tne aren ious ose the assoc tation suse, as “customesis whe lous byegd also ® Lendl to iy ute @ Clustesimg 3 “This functionality ts used Jo gorau the data in. tancet into clusters 0% “Tic Cunctionolits 1s + ____used to Mentley ine Wate —__—|nstances Arat deviate s ntficantty form Me ——Horunal ov expected behavior - fart example , —, outtien cmalysis canbe ised do detect Hie Termeni he ete Ae , Such as _Cnecllt Cagid Porauiel, nebwonk jntusion O91 | medical diagnosis see 8 Boolian analyte 3 Ts Activa iy ts esol de analyze tue changes 09 7 itsends In dno dato ove Aime , For example ewalution analysis can be used to mondo dhe ,porrfosumance Fo business, the behavion of a_ ——}-Customesi., oot the development of adisease. —e— — > Deacsdbe ahout Major issue tn Data Mining 15 PRS Dota waiving ig dhe process of extracting useful | Information ~gnem lasige and complex dataset. __-Howevess data velning caces many challenges ond IL t es uch as ' zee | : ___ 3 Dato | uality > The way be inacunate IL f > Inceasistent on eis which can Affect duo mesulis of Dato vu qs Data_cleaning. amnd_porepaiecessing techniques nt used fo Improve the quality of the date. _—__- \reacher's Signature Bee PE eee lla2) Dato complexity & the data may ied —Vasdour sousteas ’ jin “and ata high Volume and velocity , uhids can winkatt « Aeet cult to process, analyze and Undesstang 3) Nata Privacy and Seuxity “The date yoy contain | gcse Pensonal, sengitive, oo confidential infosmatien Jat must be PModecte) from_ unartnosdzed access or mistso » Data privacy xequiadin —T. * __ impose Sted of suiles on how data can be colledeal, Used ____ land shamed . ® Scalability DD Dota mining algosdthinr mut be ie : ables to handle lange and dynamic _ E | dodageks eteldeutty amd @Feoctively + parallel, distodbuted, and ficgemental tied al Food trams age. | Used do neduce dro twe anol ¢ putational recouses {equi pred fox data vdlning — : _-) Why do We prepocese Hie data) Qiccu? UU + $ NEP NOCAS< | ial ctep in she a dasa analysts precess. T+ Involves clean sans to make 4+ sultal =i eee cede Data Cleaning > Row dato. can have missing : inconsistent yaluos , as well as nedendart eee cleaning _nyolves, deudr eying 7 ett—, Connect fhoese. eornoves os: incondctoncies such as ptisst 4 Values, Oublions and duplicates 2) Date lta raHon > “Dis inyalves combining deh ale from Leet Lipde Soupices to (Heat __.a unified aba wit dui Henert __,formods , Gtowchusies and comantics can be-chal 4 Soke. “Trancfosvmation > Tui Involves C the date {ro a suitable | [feeumake fos analysis. common teciniquice vse In data. transfor motion include navunali zation Stondaredization, and discoretization. ___ af 00 S Reduction > This involves convessting dhe te dota sizeof ne dataset colle prescowing to [upeotand (Néosrmation -Tedwiques. [buch as Flatune celectio and featume extraction Con be sel foo dora eduction. — Ft laste detais nore on the follousing < ie a. object Odented Database _b Spatial Dota. C+ Tevyostal and Time Ses os date. di _Hetenogencous data. __mM dhat can Loose old Compal ¢ Ck cata object, Vas objects nee aes ae Imebje para enteal daft ‘ = nes infor In_fhe ~and classe: l Isem 0: eqtunes [ike igane ral A A_Spinal date > spisral datu fs a type of dato Jat = follows a spinal pattesn om shape . splsat data can be found in Yasious nabusiat and = physical domarns , such ay DNB Cyclotmens , and ma + SPI9t | paremetsic eguation of ine foum ~= a colt) ye cant Yeasitt), 2=b+ whe @ and b ame constant and + is dhe: penametese El Tempe d Time gexdee dota 5 + ie : Sexier data is a cass Of models foo swe. esses date that can exhibit near. Jong-nange dependence 09+ peas unkt goat behavies.. Fe. pened He sextes data, modietes Sele Lous Kes of the iv Jel by add) exponential tompoodng ae The Jou i jos f 10 BL that pat nonctend mates of Concistenc mosumalizati aie Hetexegeneout dlata = Hedenoyencous. asta is date Wat lnas hig hi ganiiabilits of data. Hees and Fivamtt:yetonoye ntsc dace anh be possibly ihambiguous and ot ue MUESINA Vatu, — paige ~ data nedundanc untowsnéathes Tt is di€ficwt to iy errr ate do meet tro bucindss infosumation demande ae mation sa $ data can he analyzed sing _ eee OLN Os data cleanit HANS Fou, data. indeg nation and data pin — eae Wedte tn hod ee about Date cleaning ¢ 2EEEIEE fs > Dota “deaning | is the PNeLess oF fixing —|f 0% s1emoulng brcommect, Convarpted , incor formatted , icate ps -incompsete data wi Shir a dataset. Data cleanyne is essential fo _ : ensusdng tbe Quality avid elakility of he data clas fhe perfomance of Une Marlune leant Wmodets trot uses Lbs hata Cleaning typicalty involves tne Fellosng stops as ab |b Revove dlupsi cote, os. Lenolevent observations _ Fix touctumat esgoxs , such as typos, Capitati oot navang. Conventions, > Filteoe onvea anted Cubticee, 0 extreme Yates E | ke deviabe From the youmat eine se Handle missing dat etre. byt vepudin oe on FI tl yoi o date, such a6 1 Cheek ing ut date ety nges om fonm ots: | Seadyorplain MMe fm) lore) TAY % pate Lrdegnatien Lata. “\nanstommasion yctynde , i? @ Data Intesaafion Lota Internation ; PTGS OF Cotmbiniy athe dota $045 datu courses nhs a _ehenert dota atone , pri uid @ Unitied ‘yeu of tue ne Aas. oF data 1-yeara Hon 1ado mare se dot clu ong) Macon tes fn the Purposes of end dedeies rigkiag Tecrniques ted iz dato Fon Srdude dota. wanchousing ETL extract, vate fonin ,\oad) prinesses , fe ane Mauls Zmajon ap) dafo. integrate ei. dhe "4 ghd otLIA appmoad," and asnenits ine" loose Coupling BPE OAL” rie wutbhose pete TO9e, - a0 ody wud $479 ©) Dake Bran: fonmation Mebaods bata Tronsfosm ation motned in date «tenveuting maw data fom analucis and _ Land eding fete oes Sparstoxmnat ian, 1, £0 prepay uke data fon dota mining so J Haat 14 corn be usecl 40 ensrad 3 reds and knowledge. Data tsans forywakiapn dae ica, Inyolvor Sevenal steps in luting 1 4d) Vasa Souning 2) tase torcafaahon Dobe peduction Lato Viocned ation Lote Pops raahan .pa eductian IS @ Porocess tat Yolume 6 of data and arena dd 10 a aes Sma, inins ¢ tf te —teedginal data ae mala} Fa of saa Alpe tim tl ce Nenhonce prrecetdns oe Heme ane sone deena Gsed_in data seduction ee eatin Discuss. im ndedaiy i Dimensienatity Redudion 4a “Tis technique Clim nates peak ly [wapo%tangs on Seduction Featuncs from the datas | tvencky edu. Volume of oxiginal date {These Wore toiee retinds of dimerdiarality aetich ® \lavelet Bransfosum « : fe ee o prin pal Component Pr S\S. & Attsdbube subset cplechod. i y Dota sampling | ~ This technique involves _ : [ ele: Q Subset of fhe date ae _to_wosck with gathese tren using the entisre jdatasel. 8) Bata ufa Comparession |= This technique Lovelves eee y8\ ng_Loss oi ldssless the size 06a ee [Compe ss for reklreds to Mecle: doterce.4) pata Disnetization *~ ‘this teumique tnvolyes af Conyes ing cond} nuous ; iporrte dada by partitioning the Mange of pos ues ro ipesivals oot bins , s] pata Cube Aggoegation - amet? clatnin 0 simplen Poot data Into St ble — peseaibe about Data disceHzahen 9 acon 2 Dota discnetizotin js a pnocese used in featum _, doansfosemation to convert continuous data {nto : | categostas dota. Tt does so bey dividing we grange of “l4we Continuous data into a cot of inkeswale. Tis procedure sransfosuns continuous yardables into | discovete Vasdables, and i+ is commonly used | [mining and date scieve pac well ag fo drain mod ele foot oWttiicdat intelligence . Pn, “| Hene ane some commen methods of data dixcsietiz~ __|l atten i— 4) Hisdognam Analysis $f plat used do siepoicsert the undesdting faequ j istointion of acontinuet dela Sot 7 smoafuiny tee hnique - -that help as pup a lange pumbese of Continues Nolues into smalles Values, aceeanaananaa — — Sh ast Pmatycts 1 AL clustouing. algositi ; ___Is evectuted by dividi : : abuas of X numbooes Into clustoxy de isolate « ______,lomputational featune of x- a ey ESSE fee See 4) Data Discnetization Using Decision Tree LPinalysis $~ Atop -dows clicins Fechnique ic _t USed + Te ts done shy oligh c supey— {| Afisea PiHocedusre . a ie 3 Data Diecuedizatien Using Cosmelation = Proalysis 8— niscoedi2ing data ba linear sea nne, SION) oO Ou Ca: 1c i post a inteswal ,and tren the large Intesuvals ane combined to reo (anges oyeddap to fosun dhe Final 29 9 infesryals. | Wodte about Dimensionality seduction retank’ — og Sct Dimensional Ue eat | fo wecluce “the number vf featunes fra | dataset while nickalyaing as asso of fie jmposdant _____| Infosunation a8 possible. TH is @ process of — — tan Lomwdens igh ~ dive: mensional date {ito a lowe __fdimension Space Hat ieee eet € essence nf the osiginal clote._ ane | Heme Are Some Common methods o: || meduchon +— Teacher's Signatur| ea 4} paindpal component nalysis “pend — eee CA is a Hg LTS eae er —_[statisticat procedure tnet Uses an osdhogonal 7 | dvansfosun obs Yations of [possi ible Cosata elated nie a Set of Values of ey peered | Components « ineas Diccru 6 (LDA) t= ; LDB is a metho: pitas in stadistict atten Decasnition and maddie loapning 40 find _o lineax combination of featuores shu! chasu S420S oO9 spepanates two 08 mone clascas of __lObects 0% events + 8) Memeralized Hiscodminanf Pmalysis (HOA) — | ADA ica non-lineag Vensisn of LOA; which [cam be used when the clases gore nek lineasely Sppanable + q) ghyqulase Value Decomposition CsvD) — VD 1s Q Factonization of a recut of tompley medrax phot genesalizer Hie Og endecomposi tion ofa Square nesmal massi dro amy Tm xn mato nia an extension of the pala docomposi tien. ‘ ') Featune SURAHM 1— Featune ¢election inyalues Selecting a Subset of the ondainal featunes tat fase woe svoleyant te tho prrebfom at hand revicod utoe ncod eg Aye a type OF as nl neunat network Used fom learning creictout code Of Input date . Teacher's Seno> Wodse detail nate on he ¢ Collowing. ee a_Lengunsut analysis Cece). ee : Both peindea Component. Analysis (pef) isa + _Stahisticat poreduge that les an ost aeammat a | danse umation. do comvert a set o€ coselared ___Nasdables Into a set of Uncosinielated Varjablas “Por is Widely usect [7 Explosrator data Lanalysis_ id —Tana in machine lease ning foot poreclictive paodels : The wan goat of pcA ts to secluce dhe olimension— ality of a dataset while poroseswing the most ——tmper font pat tepunt as gielationchips hotween te Yasdables wi fsout Ta Pod 09 Knowledge oe wathe Sasgot Vatiabhpor , (© Facton pnatisis isa statietical teclnique _ 4, Lseot Jo mecluce alarge Vasiable jruto a smaller — Vamabie factor 74 eyforacts pre May imum Comma, wiance fs10m all the yasdables and_puts Shem _ Linto Oo (Common scone, Factos analysis isa past 1 Ot the Crenesal Linea todet (Lm) and t+ pelleva Sevettiat preores shat Contain mo Mull tollinea outs | Lin melakonchip, 49ae ConhelaHen and vouh vaddablor tno dic anady sis aug face and Yaxjablos. Teacher's Signaturethat 9 you, YAASIS tony dota _Maisdite nakes on dhe following .- - pe GY Disestote Fousdesc * a ~~ ® Discorere cosine © Discnete Way co ™pnession 9 Mestoure Bow en OFT) @ Disesvete Fobsuiest Totans tesmatten %— Thepi cote, —t i conyooss a Finite Seguonve 6 xg, si FB Data Grenasim 1 2D) Data Compnessien Is Fhe process oF Ysin ending . and athe mocdiercationg +0 xéduce 7 she sire of igisat dota Fier Without enateg Sheise reducing dhe s} ot Files | fousuess Trans fooum (DET Is a matematical tecknique Useal_in Signal process} otnooe field. Ty Jor Spaced Samples samplod of tho digcowto -+ime, OF & function {ndo a same-longph equrencl of emul ouster tansfooum (DTRT) whidh is a. Complex-valued) function of fmequendy __(b) Discrete Cosine Tranfosem (per) &— “The Disco Coste Tran form }s a special forum of Discoete 7 —_||Fousd ¢ Ree Tnae ubationally Ughdeos —_| eal in speech arf mage | 31 ny tompoasten {Coimporossfon hecause 04 — ac Sneve ompartion ‘ia Teacher's Signature Is_aseel 1! En Loss, {mage! a© Disesete wavelet Transform, (DWT) 1 The Discrete Wavalot Transform, (DWT) is an, wavelet trangtnoun for which the wavelets asic |di.soo od, AS With abner wavelos romans | a Key advantage it has Over Fou, fistamsfovums ic tevepooral nefelubfon ft capture {ooin fre uorty and location infomation cation in dime) . DwiT decomposes @ sianel into a. Set of pustuatly oathogonal wauelet basis €uncttons. oanelouse as defined ag dot by willie yy OE data toanehouring , defined a date Immon, the fathen of data warehousing, — | st out due fous major features of data \Wil tom In won, neteanized as the (olue || wanchoure wi tke “following foupe wayose Chanactosdctice nana Subject-osdented !— A dota. \anehoure te es : O%ganizedd Onound subeck OA mayor. heat OF heres’, water bnen fhe wayo% Application oF Lue oxsanizahon FrKam~los of subject nude cales/distedlul footketing ote. aaies — Teacher'sA data wanehouse integoates data, data from multiples sources y 5 buch as “pyalnEoiames nelat ional databases, flat Filerete, ae Sc Se PERT OPER FETS Sougy t7 “The data stosed 1 wane house prevides information from a ical penepective and is associated with a _specitic dime pesdod, This ig mone extensive Compare to othose openaking systen ai Non-Volatile $— ta non-Volatite data. warehouse, data js pesumanent When mew date {s insesded , prentous date je not joveplaced , omitted og deleted «Tn this data. wanehouse, data | send only and only sefsesher 1 oF Cobain intonvals * Dekine Data wanehouge F Discuss Design ipindples. > 1) Pata wanehouse [5 @ centralizes repo “| System where business 0g Stome ane| process (ase amounts ef data {sem “Masoud soustec.s eat business intolijenca. ang Qnaty.t ies. Lk Publ ——— pagettens data. Bue many dd -ditcesrent “soustces intonelpics and decisien support. Tk stoos ut sinical date ovens dns and enables {ques} and Seansfoxemation, date usinehouse (scr campher > [designing o [praclsc hint qaryivcas cometh fadherence to a fem key podncipl { __ in a dato warchouse Fs described a: slanted « Hmee-* yardiomt non =Y8 intogriated cata nepositeray fax Me ey endbepodse » : at = a Poltor= Up Dosign Appreack t— — ae Tn dhs apps (@_data veanehhouse ts descsited 05 "0 copy 9 ‘foansa.ckien data. specific azchiteckine fox gu Tord craic # teamed tne, Stans ccleewna, @ Nate t— 14 te impostant a0 slew dine be 06 a data Wnanehouse do youre bu (ctakeholdern Vey easy on tn sve pele See ee nad euing ord tates (eto maintenance (ott. ——— Gi ndapiabuity += The ability 40 adapt so evolving Mee eee S fousimesc demands is coussias Sponsor contains a.sot of eiolated facts This made je byelcaiy used n online Bmatyaical priecessing (aLne Ing. Trovegstesends data in Fie feo ow modeling and Viewing deta ‘and _poxspedtives. abtidivensional date [Lancl_da 1 oc deri canbe toh 4ypes et schewan wedels Ie “The simplent tyce oF ccbemat | Tt includes one fact lable. mfesenciy jag umber of diveenlon denied: The ft Kable Contains Keys do Cock of Hye climentions and alte cludes wiacunes, facts on metrics ociated a @ - The Success ject te now mensut Phot wet id cvsporsens the business. Ido ertoiact value out oC the Sycteen. Hee Self -sesuice BI areal olitionsion a [Snewrtake schema one comple aire i Some simengin Antes [ane nermabized, Luenohyy (arte splitting tee [date indo addidional ton'as. {is schema ste ©I } 4 anel mabetalns dat intern ba Seducing sedancancg. cae Jota de standard 9elat Shooto his Schema canbe Sern as @ Coltedion et a Cohowens Ond is mone eon play Tt allauss fore Tinudtipte dich tables dn shene dimension Jabka } “Top = Tew: sic te Cone nd trae + mucstin no = 39. [prepending deal ais ond aad 3 2 Er plain ahaut due “Tree - tex, data carie— Neuse amelaitectune ushtu a neat digaoam ! > coms giana aarie| | ee BSP conenuatiy a dato wamchouses adept a fhoee | [eyenbreres os : Persians cprareresitscaee reer { ‘4 i Honr cf date yumchouse arhitectune.— | (ue following, diogzam dep Janchitectuvie of data warche 4) Potton Ties = “The laodtom sien oc the aunckidedtume is dhe dete were “datubale scxyes. 34 ie tse saelational date ba Foon style We the back end tools and Ohi “peideoum Hue Exdnact, clean Loach and se SS riddle tiew — Im sue middie Vem we have _ ft dhe olfp Senverr that can be Teplemented fp eit en of be fellessing! tanya LP Pdatignal ot fe (POLAP), wide ic on touted Lanclatizsal dntabere args cment sy-deen. Tos TP SURy wags 6 operable’ an mistress| Diseuss Fhe Folloia ——_ 3 star schema, SNow flare ccuema ~ Fact conste\iata, Schema, - os @ anthem. a Llation schema > tnt encieltatnn | —_sclatria tn regave seg |pruttidimesteral made 2¢ fs cotiertien at wag iat Aoi faving. Same Camomen dime sin Sas 3 cab ew ‘callechimo€ several @ pase cubeid: Cotudent cot ; — ___insteuactoss, count, oe ~ ppaneoP lara tir seer eecced __Jeanear clown each dime 4 asi = ables fos Cxamples a ie — BL OURP operations Fe tia Avenaze opade ofr —[euises fos cack students a>) gedit down, ee stadd with the base cuboid Student, (ousue, ~ semesteny;fretoctor cl» * [Dadi) down jo Cstudent, cousue, semestead srevmovina tine Instgutctom dimension Teiitoas 16 mnsults 40 bnchade: ony. ca Poll op siowt wit tne pase cauboid Estucent cours “SemestestIngloutctoos} . 2) Poll op Xv1am semesten to ean by neerov Sroeoien climerion = Taille bie sicsalt da include ents { stice and vice se] stant with dle base cabsld Ctudent, course on cs founie Semesters; Inctuacler lice fhe coke bo focus on) [Bice Mine teabe do Gacus en specific. students [neces t s@uestion Denn 24 > [suppose Aviat a data wanchouse. consti the Shaee alimensions sme ,dectoss a [paticnt and Hae dwo measures Coane. charge, whee change ty the Gee took, dectom changes a patient Cora visht aise dierenent classes pf schemas +at | popalaniy used Com modeling dota a meleas ® | solect A schema with ’yustt Pica tan to Tne teove teomeliause em {= I how tne sclnann disgnom selertol Ini fn foe above dads inelonae pos/> @.pifeanant assser o¢ cchemas poplar (Bted dons madelling late woonehenser includ? ‘his schema design has acentrat ie_dimensian tables Hhaieugh: Stan cehema act onsen ¢ a ee et nach trefan token visualized with fact tables at Hoe center | ond dimensions suadiating out. @ smowe ne milan do the stage chem but the dimensions 4o1¢ Nosialized into multiple Iroladed tablet. corming a shape that aescrsbies f@ snowflake. Tais can lead Jo mane e€trcient stage lbut may complicate quescies . A} rataxy schema i An exdension of ihe __._ schema, invelving matt tohies shardng dimension \able< 14 is suitable fo re complen aind divense date warehouse we yire outs Fact constollatian schema - jnysiyos jmultipte | Cact tables chasing dimension dablec. This ts r= Complex Suen swe Nar gchiesta and js sublabte foo [vor iaoge and compeer data mariehouse sceneniss,—@_sdlectin of echema — [rom the given date smnelowe sits |. Chime, doctor, patient) ond measus Live becuse the stow schema is ci Jundoststond -and provides efeidett, quer TL pemeouenance «The dact table cam cabin “Pkeacumes (count a “Tables Cline, doctor, pahieut) Can be disvec Connected 40 the fack Aable _cdaeye : ith hustie: stan schema isa surteble cl change) -and the digensi Schewa Diggnam fon tne selected sia Fn Mais a Tpockos” and “patie” tobie ane _| jmensisn dables How padnaty-Feclpe Fey alot ion 5) Enumenate dunce cl. asset st asliciet Nook one pianig_osee foo, mndellg tie see gers qourchouse conails of me - —Tecttewing fours afetensians datn , spectator, location TR game rad 4u00 measures count \ clerge. where _ ibege fatie fare spectator pays when twats - Tike gime ana given slates eee ay be student” —Tralalles an seit iri ewn-charige orate eau a cchema diagnam tar Fue above Jara S warehouse Oring suitable schema. @ Tutieg fhe ose be Fi schema vied in Ca). GAL Thnee poy ia: Laemias sent 0% imadeting data worehouces are! Tna stan schoma,a centou —_€act table is connected to multiple I Tastes Conslgn. Keifer eh ‘Fite. cact fake contetes then crs in aed Sealant ti ica ati, nian eerie atodputes, osdan sche) Similan ae.a stan sche @ snows Inte sclera. alee fis centnar L fact tania a abet the diwensisn tables can _be nonmatized , me J eed meaning 4 tox gre Custer ctivided into Sane dimensians Seer doa. ciructune {PiepEMbing a cnavoflake pritte the fact lable tn.lane Condon and the nom mnadized di mensions i bmanching, oot - — _ et alaysy ccvema (Constellation rakiar Bema) Ha) Aolary schorna {c an extension of the, on o£ the sta ochoma cowene multiple fact tables ase Cenc: ck jbo multiple. i nei; lolso Enon os o.constellabien cchema becaquce __|ft mesembles a congtetatton in iis layout. heya. Diagorayn = [ = : _ Tract | = vasyie L } _ 3 P epoaareT] cation [name | 2 Lem bia p GY) Sushi ticaton | — eeimplaty Easy do Understand and navi ©) GueNy Penfommance S Faster quenies due to dutieat Flog = i flexibility 3 Altows foo easy expansion and ‘modification i el Usen- friendly = pell sited coor analyt a reporting and analyst