0% found this document useful (0 votes)
11 views16 pages

Unit 3 DS

The document discusses various predictive modeling techniques, including linear regression, decision trees, and neural networks, focusing on their applications in predicting values based on input variables. It elaborates on evaluation metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), precision, and recall to assess model performance. Additionally, it highlights the importance of understanding bias and variance in model training and the implications of overfitting and underfitting.

Uploaded by

sudeshverma2648
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
11 views16 pages

Unit 3 DS

The document discusses various predictive modeling techniques, including linear regression, decision trees, and neural networks, focusing on their applications in predicting values based on input variables. It elaborates on evaluation metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), precision, and recall to assess model performance. Additionally, it highlights the importance of understanding bias and variance in model training and the implications of overfitting and underfitting.

Uploaded by

sudeshverma2648
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 16
— “— UNIT- WT. Date, = L adel fing :> ii hoo 24 ro isha) dale to jhe TA —restnglll fella 2nd. “Theale! gp ine We, an d C wm 4 | he 2 oe he shad it_wight Le a tajaklind Shectic model fo a dala “$e Called Wedicve Medel n= 7 = =i a ne mt }o__redicl certain yolue using Yoo values in an i = The a meclol allen} 4o_osletlish a _velo}is eolive je ve being teams Hs eadilue nadele bn’ a* len fn hy tuon) lean and thaw Yaey Lio} fo leas evs iy “Gea ing ia|loss in a Cyickol satel. ls ay Gedicling tuihcthor @ — dransacbion ii graud- B iy Ladies tuhethe: a _Cuslernes may rove }o _analhon Joreducl t f “The adel: iuhich ar wed Hee fecal o| Jeage) 4eoluves af Ay Colesevical value aie Krowr as sificalion mndels. the ael _jealure H Known jes ke. hick rlauer ove _olivided sae. ace called le vole ae of © Yr fabulor clowilica}ion models includes i= neares} Neto a Meigbiaws Mod), Naive Tayer and Decitinn Tree The senndels hich Ww ichion af} nicl © Malus of Yor Aeolure a] a dole _insjance aie Known ~t) veysevine vandels tines § Regen and Lngislic Keavesrjnn j ———1madelz aye ebulas Oyo mion ode a Yall. ceesevsvvvvervvvnn SE stiwaling A uachiont=- —Lalisnlig o—funclins yb ieall, vaeans sbi vn nh Yuinble L_ autul miable Y. ~The Qlimdled fun te helps ppredend nr _vew, neon Valus ox Me 4. Desi ne do alolon to YYou went fo etme a duncline } sch 2 ivf deduce Linda)ondon\ variables) sy: | lon). _Vaaiahles) 7 lenge] _(debend: = error _}orwa/ noise) Depending on Your alae avd geal, vos orond tyber ol dumelinns * +... + PiXo 3 Wain ho oadal UFsVivnale he ara veel ):> = ae ! oitimale ths Like Han} Syn ‘machi els, you olen use ablieixalvs= Mhm—4ovninieetze bredichine ake wnadal 1 evvor. Lue wnelricoy Like : oon Souare Frrac (MSE) * a Spuaie ccurag Qeciston rf Remi ole. daponding an_18mewion ae pCO ov ica pions cers Sgn i ‘ar oxen: 7 tak ting. house oe 4 . Talal aa “isn't Sela cael ae orlunioae Algovithmns Corarnnsaly bat + linear 2 + Decision “Wee — * Ronda ~Favox 2 + SVR ([Sublar} Vorlor venor) A + Nowal Neduork’s % => cation 3 oi Ghose: Validea Cited label si CO \s 'screlo ‘claus lex. 13 Shave or _not sham besidive or snegative ) Evomble Uso Cues! t Eyeasl show dobection Algexibbwes " Ysed + oe < . . Legis Regue sion aK ona Ox >: auipier « dvm «Kk WW + Nevral Nols old, APCO wr Fealuves Rogrorion °C ica\ien 4 Cnthul “ye ——_Cnofaunies Ve Duciele class /Gleqacy 2) Evaluabion Melrive, MSE, RMSE MAE Accuacs precision : RE Score Recall “F4- Pease 2 Quotion tuba} ill be the ill it vai devalpreolwe fornnor rou? tororrow. an Salle vn ice wed onto diel re Vosilive Negative Qasilve “TP p Nepalive En. TW. 4 Wr fos ive (TP)! “The arlual ond predicted values au 9) “Fae Qsitive EP): “Tho aclual value aw dalro lou} the feel a 2s Hus. “fe. Bevel: =e he mali au 4h “foe uke aclual clors ot the dele oie! tos rus and the \pcelicled iy Jake. up fae A Acourocy 3% Accurary af brodichine is hea C ae Clases z sian face wis vs Gyrectly the Cle - S jam Thal is il shoul Cavselly clansdy “Tes 9 1 Tobe on Jab : ‘ Aecrnio a CP+av) = APCO (TPs FI TN) Dobe csinsisssocrivers re Lrecision 2 “TF is a eneose tells us thal _lprebrer}i tains! all hosidive Sabor, Fin Soe yale ise shu led ecision = —TP (TP FP) Recall oe Sensi}ily 3 A} Berl! Taliale ‘ hala of | Cover! brediclion od hosilives to ho Jnlal ura hen ol” Lnoi ETP FN) “Ty _is_a edna Jal Yells vas wha} Sra peion xi o) _Guthu) clones which ihe aw _\wedicled 2 be. =" indintes ged lowe 2 -ynaild being —einauives "i Comeryalive or xcouivel age ecitioty . “TW a: ETE) “FE eneonne > “Th Gu bines oabines LH, hoe Cision and reroll. “Jt torer Le ream of precision and recall ay ZGaloulated ax: APCO ‘Teacher's Sign... The geal I Sym ‘Teng hia and variance. A Diath. oravsonisiiisenn sh Wes ih KD) Bo | [ESN (1 Fo \\ Rte Coy) =o 4 low Varinne alt Lp Yosionce (Suality o Fil € “Tis 0 coial apecl of slalislical snadel ad dela geass, “IT bulbs fo dicing haw ell 0 yaad exploia Ys “cbvel_dala t R. 8 i jon ) 2+ ¢ yp xehrerents he bro bor iow 4 Variance in Yee deendend vmicble tal Gn be eine iy the inden aichle, The madd APCO Teacher's $i n Gable seecvvvieerrevvnne Kara, w ol i ( Cons no erblanatecy jouer. 4 rae jade: 4) ws C be vaisleading S8yop Wine ZlysG¥ = Resdial Sum al Spence (eserc) SSpaj = als. iy’ = = Tela) Sum al Souaies “Tananline te beer, sb Uvelictions @rac}hy male aclual ; She aclaal values) i ee QL: ) Vahyo) Q’- 0 + Medel brolicls na helen Yoon tho aga Qi “DH Caleulate Pho average Syuaree) diljerence bekueon he aclual values ond “tho bredicled val Spuoring em\haizes. Larges rare yagre Low Syaller Ones —Farvaula > mgr 2 1 Zly-fy » o ds TTajer|arelchinn ie * Smaller MEE eons tho vordel’s ocedicbions are es Closer fn aclual Values $ Unis ais _Sguared, which sigh} nat be inbuilive i Pi aot M lean & au Exon i> a id! iv Vv’ 0 lease hao ag segs +4 Pe d 2 “ewore helueon bcelidel an) abienel Value Sgicing 8 APCO that's Son. Dish csrsevvvvennieess of how Ju a} facedic}iows ave Jaca ac ual ou).Grmes | Facroula Quse = 1M9e _—Tejenrela ion > * Lone vale mn Indice the belle dif. erbre hecaus if isi $f Moon Absalste Ereac (MAE) MAE is the average od Yh absolu}e diMeren Lelueen loreal Hons and achual value. W Unlike MSE, ib Ieeale all execs gually (no spuating Favvula > La g gd By MAE = Ju ~ Gel Bele tt : Pace—sahuah Je cullen Yon MSE. Bete... Sirahle Vinegar xogvemion > — Ono. codiclor. “The ynodel aurea he 0 nas Vario d Par esins mm Linear ew \ho dohonds = Vphendent Varidhle ¥ as Y- jnlercolt be Slabe of te bine Ye - =e dsb endout Vnniable Tre Ee px the —dabndaa!” vaniale and Ys cea a Yhe Uaetly da hs }rodiclar Vaaine, we ean build a ynadel wag dimple —lincos Reith v fice 4. ree = 4 {Area pepariy ) tug Can velecroulale Yo voodol Os %, Te poh = a eb Area, Sho Saige} bine vespectively . APCO ‘Teacker's Sign. - ou tp y = © a ine. inleceh) Jove Y= ath hese a | o| Shai tiv. fi ee eel) oy labe ol a bral ht bia thames Ei haw much Yo lina in a avalh change athe Veelical alirection Diab rR | Fa yes | - ise i Raw Neb rerenatiow lohe = v2-Y4 X2-x4 “Then Can he Jue Lybes of leben ix Linens coy vei sitive ebe ay negative globe HHedo Types 0 ad wdinet ihe Job. + Give liner osibive slelse * Linea negative —tlahe * Cnye lincas walt globe ee dele ale fm 9 grahhs tye {Sori hh : i APCO . x fines: sitive slope Shes Rue/un = Ye-% = Dalal 1oehale) = AY XX © Scenaris 1 jee pester Ihe dallaly) js sve and fella (x) is + in 2 doc }osilive_s) be: Dellel y) is vi avd Dellal¥) is -ve. Cuwe Linear osilive Dobe: =ve_slebe almoys raves downward oma __arabh rem Leds Socials} » A. value on X- avis) inerene 4 lus dercenses —¥ pe APCO x Dbessisasarcasce alebe= Rise) pun 2 4-¥ Vy oy = Delta Ly / DalJacx) + Scenacia 1 Jor nogalive slobse? Cilla (Y) te ave and Della (x) ig —ve. Scene! Scenario 2 Jer negaliv lope: Delle y) is —ve and Dolley) is +ve. ve bi v > | Va relotiomb loaf Evroc in Smale Re rossion 2: oe Leawisg tase the aleve does ileie| dermal ia algorhae Y values ox lprevided ty Yoo machine and i} folondahn ajo ticlesrht) ond bl slob hy —xelaling Yko Values a and Vy. Mewever, idev dying {no _exac} vaalch od u APCO Teach’ Sigh a bossib The | rn Yo é) anrocia Ht This_evror_is alla) rnarainal or xeridua Cxvor.. Ye larbyae Sta}se Lb): be ngxy ~/2x) ley) nZx2_ Gx? “Trfesteh,) (a) a= ye bEéx n Dahle Tinear ‘hegvenston 1 - a Ta a__ultible reg Dod & © WTR oc vance tadobenden) varialies Ove involved in the weed “TY abou, dor a nore Combly onalsis of bow t t v daclars ind ute an outcome. : “TE tus “Comaider fico al Th ely ar th friclo and Aveo | gS hion dlear. wean Sine fuachas aud Geaoniiex available av dehoriclo ‘ariab! es pm 2 . te egartg = 4 [rea oy sacelinn, Alone Aashace 7" ‘ S Aveoni lien ) imble a. yaul till “Leareuio ws male) ayer tho arinble i on i 1H now WZ, Ve a+ bX, + bX “Tho vnadel aloorihe @ plone tn Poe 3ZD Shace ol YX, aud 4 —‘eaarachor '0" | Yho jnbercelo} of thas | tone. ro ler _"h! and hj’ oe ve} ened ‘ta af Yulial iene APCO S Teacher's Sig Coetlici eoks Vuowechen s xebres ds Heo Longo tn the sna» _ vesbons Corvesbanding So. o_uni) wae _in whon Yo} ‘held Cons an}. : e ‘Yaromeler ‘by sehrere ds Mo change in She mean ‘es nieshondig joo timid chase in YX, uth ey Conlon)“ Mu lsh cepreusice Jos ea}immaling yuan thon thor que in! _\eredicley vasigbles_j Y= a+by, +b bie ot beXy +t ba Xn r mile dining Yoo heal Hh lise. tue Can Jl either ably nonsl ax_Caruiliaess co tte ii oe labaonial or __Curyilinoas coptui xeaberhive, Assiaminea is ‘ibssonin Aralysn > 2 ~The na, of obsenali fn) is 9reater: Yan tho no. el fearoradon ue Ye be imaled. wrk. 7 Weloiew hip deb eriaed ‘oy regvension ars daly _velabinns his © a} amatialion ‘bnsed oma cal sel_auk nol _necersaily a) Cone 4 2tfecl ol the dojined olan 4 Saucy jon live Con be fall oly over APCO Teacher's Sign. ange eta. be vo hurmon Condih'n ange andthe purines ars ndery ho «earerion neds) aie sno lp alid how Pro Vala gel tuill no longer be able te predicl §) Variance is Ye Sawa fae all values oJ 2 ik mocwally dis ethjal- & is inddsondon! and a nol yelaled le @ny value ol) = dineniioralily veluchinn larelirlaes (2) hea penal and Mno node! is go). ing tho trowtor- : ooljer dimeninnality vel “Tien. OIEZ vaure bet ans is x2 Me Laing Yho dimonsionalil, veduclion. emai! hone analy si 's one _@ a) diseenscaaliy Ceduclina — hochwia wes ghimalian tho Coed icie * h pi 02. ot ba 9 a Teaches Sign

You might also like