s\\F Tntroduction to daa. wayehousing and tala Mining . Pala Minivig Ayptem vequives imput of cota. fiom ‘Three types of storage. Environments. WW File. aysien (®) DBMS CPolabare Management sytem) CoLTP) 4 S Palawarehouse COLAPi sytem)’ yi od W File System: eur 9 Th File System jth datas Organized “with a” set! of files. Generally in mining system the Input & only given rough a ad files such as + CSV 5 +HLbs ranfh ete open DalaBane Gmnechivitg, [Some cl tyson (ariiitute velated file “format ) iP Avadus ex ove. palatine Connectivity Vorlucs c- oujec J inking embeddings palabase (cormeck ity, ar > The main cdrawmack of a file sytem ww bins Seely , no valideiion on a data, lock °5 management operations ono Dara, | sosat [luge @ DBMS * | —> Tn DEMS, The dala % orgavized with a Sek of DalaBare, Generally Dajabares or wed do Hove Lage amount of Gperationol para: -> The main annnnges of DBMS Are Nigh secunity , ‘gt fetes validaions —cha. "aks ehbickenk iatonage mer = operahons ono pola etc-,b y ~ (3) Delaware house — On Odlawnrehoure » The data, Qigonized with a different ounces of daka- thot means it con stove hebra geneous (ip, Catection of data. it =a Generally Thlaware houses ave Used for storing nion-operaty, or historical data , Henig ora “The input of data requires 1 types of Prepacessing acti Tuy anv 7 i20ala honing ~ feriove Ae -neisy on dala, ill the miss Cn ‘ “ality Rermoving intomislancy on-deta. ‘ *avUrfoyrnation ~"This i the process. of Conventiing the dota into a suitable form for mining applications. (3) Daa. Integration ~ procs of Combining multiple date Aources into Armgle Paka source. © Ohta Leduction - he Joige Volume of data & comet into smoll sek of veprosentaigns withou Lossing any in-ermation, — UNIT: Tahoduction Jo "bake ting. (0 Whats Dlaining_ 9 origin ‘end, ohana challenges (lenaualedge piscavenss tent , GcDD) in elabane) . ! i) ae a vil : 1 a0 ee : P Semeigh D0o'o cheney eolepaltenytvalvalion —* ¢ 0 MN ata, Wow four i oy (1) knowledge regnesetalion Von Bo heey Cvisintigation Techniaw s) /palamining is a proces to discover she (mowledge om large amount Of dclavases Cr Dokwarehouses or files . Tu olhen name o: taiaming is. also called as "knowledge, Discavery im Databases fe (KDD) “He above, diagram, shows the working, prows of a palaminin which indudus both preprocessing amd “fot processirg steps “Ty ane (Dela clearing ~ Remove the noisy Qf dota » fin jhe 'otissing Naluss Yempve, incimsistamoy om” doko. pala Tawsformation - To convert the dala io “suttabte | form ss aes for mining process. @ data Tntearation - To Combine the dake rom rrubiiple, Sources ' into single. Aounce« ) paka Reduckion - sarge’ volume of data we Arnall, 4c OF vepresentations: ; (S) pallom evaluakion,—- Once the knowledge i, extracted where the knowledge bs evaluaking ‘by using . interesting pattems nt " _ © knowledge gepresentakion ~ te ‘Kngusledge i presented” by wing diffewnt Visualization Techniques + puch as Tables» Cross tales , Decision Trees » Graphs» pie chovts ten, The following ave the paric origin. Of o balamining system = AS ctakisiical appyeaches : (ey Asdigcdol tnkelligene Haun on Aarne >" Madkinastearting wor wvoay ok gevibwaai: (4) Peep leaning. wth . - s tSfdilowrs fovea}: 5), pathern igakions i ; isin sant nA JB A Ci converting mo ‘ Sdtu following, ant the ‘motivational Chidllenges ofa ° i] Datatvining: system ( Scalability ~ Handling Jorge amount of data. ) High Virewsionality - Handling large set of dimension, © Complex Data, > which imeludes mmulimedia. data’ —— fadak dota web data - (Oa Dishibation,~ Guileting the dota from differnt { Geographical Jovations- (©) non Traditional stabitical amalisis ~ vohich includes Reqgressim , Bayesiam approach » Mypothusis Test pouadignus ete.” 1. Dota Mining “aks (on) functionalities : There axe ,Tu0’ Gakegories of Data Mining Tasks: 0) Descriptive , Pala. mining Tasks. @ Predictive pata mining Tasks: © Descriptive doa mining Tasks ~ Focus on feng human leprae on data. @) Predictive fala minitng Works. fous on extinatun a unknown dala. or fehne date boyd on cuonwil dala 0 | Scoring 40 above salamat Tea ‘functionalities Mis ) Elrlare) qu i ( Aa hwvobtation, ilgfumnren: log (2d Gs pallens describinga : : (ad CLasiificaluowe * ; _ } custeing . up outlier Analysts - 1) Association Analysis : , Assotiotion, Analysts monly used for mmankst Bosak analy: Tis analy \ u used To find’ the frequunt pallens from Int given dota set + Agvourdiivi ons “b the prs of 7 piscover the assce' alien vl Tok. shales a “ae oe CUCtLLs freojently dogetliin Hal! ii wT Geurrolly, auouatiom Yulyy awe dowribed fa he 791 4ormat 0f = AB tne expandable formal of, above “alte Ayaan Ne~An > BNA. .8n ge hae Cote) a saaenont atudunt => BiLys (Ce, lepiop) cpuanisation : seit | Torte classifi uv “the “process do denive the model used gush “We doka wilh a Set Of dawes For ES whose Ao duscnibe- dist’ tre purqo, OF VN ab 0 predict “the * han 04% Obje ors Wed & unRnowy" B® wutenty” ‘ ausering | ‘a Ane” coxs"4o'! amaly2¢ the data seck whidn withoul Gshy Cauuttig Ties 4 NB? apie AG «pa S wa Ut ee DLE ROE 48,9 Sothis ma apr, te iv Cu gi WT: Too principles - ha mai nab e ers breeg ™ P B 5 maxi) sing! Ya a rtiaicor ete!ane minimising the inter class ‘arttastilyy ‘between whe objet Wek MONE > The objects aw forming based “on Aira sui, dissimilanty: (4 Ouldien Analysis Oude Avolyi salto coll as oudlien | mining: or oul . by Anomoly mining: Genevally owlier Analisis u we 4 idurdtipying the mistthaviow oF dale oljed s in dala Boy. , Mt applicaliows. includes credit. cod fection, tabs imcbusia detection. « WM - Types of data - Ud Types -of’ Athibutes : Atinbuke © » ahJiwspp Qualitative -f 4 ; ; - aueriiatve ; yd v “bt sy Worinal ordinal pin Numenic piscrete: Continuous _ a - Ce + Summetiic Assymmebic ‘nkenval . “Ratio : ee HP Stole. Stabe Q) Qualitative : @ romrinal,: powrimnal., oltiloules ave olso called. _ categorical athibutes where fe, oubibute,. yale: ere defined iw igs 7 a inte re if Suk i aoe tele J, a ee hd ahs te a paseo ie oe alaiwriry Men |——— wacominghul order, sto! : exs profession , hain Colowt >, deparcyent Sees Sail ct y (h ovdinal:: The ris @ subsd.- of, nominal, attribuke and where ane VOlUEs C202" O10: reprenpnited in diffeunk categorie « b are ool Volos have, a maeorninghu Orde + Ko eda» Avink, se) AQe credit Rating. he ce) Binary = Whanes tt obi valli ave omy foo Oulawus - i ‘ wx: Gendut », Kusull + monital salu » Tossing © coin ete» TAL OM fun types. of Binary ale bukes -“W ane ~ Gy Syiammnbric. Were shy cuca of two Valles have _ occ ns eaga) tmpottomce Om priowty . | mae ; wy OE Gender a a i) Asyamebn © Wlhwre The ovkcomt Of 4wo valus mot Wate oye & quol importance ‘ 07 priority. we Raut _ masittal status ; medical foeceaet diagonlsis ports ges st (2) Qualitative : WOOLY IC | tage ME (o> Numuatc + via olin Rapin. bygone ve iy Trdenval Scale * where the alfribuke Walutes have a bof ta pal categorical nature » the values Placing oI. muperiy od bat eeu sting 9 CTL, re iO ~ Hegre Ne ve no rae ero points Defi aed alate gifaratse 99), cons F Ee Yernparar Ww “too MRS 2,Colundvr dates , stlisfactiy levels ek * Wh Ratko” seals * TY Wd iSubset’ Of Trtenval' scale ‘abbiby "These. athibucte values aio have a “ @alagerical value Paldmungful. orden, : equal pacing . bout Tene Bo True" Pen / “yoink: ex: “Fennel. in’ Kalvin, out mous, weight ee» | 6) aiscatte | Discrete odvibute Valuies are pelle na Finite or Countable PROBL where the ‘athibute’ valiies ave! -olefinid in A concrete dpprocich. thai treans te abbrlute valu Ore defined In a’ specifet ov fied, sét of values. ex: Number of childven > Ho. of Bouts vootking , No- oof lomquages,Jpeaking Cer! €) Covdlinuaus : ‘This. abbibute 8 “opposite ‘to descutt alfribute- whow the valuss ane defined. nw infinite but the values Or Voried accovcling 4o opplicolim of dota Processing. ox Height wught ; age eke Gi Types G doings) there. avs. diffeunl. types of ania ss re | ony 5 | ne ae pmo 2) ety (a Tiguaclgnal bel” mre} toh hes Ke | 4 vs 1 age!