50, willbe treated as malpractice. on the remaining blank pages. tions written eg, 42+8 evaluator and /or ea ‘your answers, compulsorily draw diagonal cross ‘identification, appeal 2. Any revealing Important Note : 1. On complet CBCSISCHENE USN [ I 17873 Seventh Semester B.E. Degree Examination, Jan./Feb. 2021 Machine Learning Time: 3 hrs. Max. Marks: 100 Note: Answer any FIVE full questions, choosing ONE full question from each module. Module-1 1a. Define machine learning. Mention five applications of machine learning. (06 Marks) . Explain concept leaming task with an example. (06 Marks) c. Apply candidate elimination algorithm and obtain the version space considering the training examples given Table Qc). Eyes | Nose | Head | Foolor | Hair? | Smile2(TC) | [Round | Triangle | Round | Purple | Yes Yes Square | Square [Square [ Green | “Yes No ‘Square | Triangle | Round | Yellow | Yes Yes Round | Triangle | Round | Green | No No ‘Square | Square [Round | Yellow | Yes Yes__| Table QI(e) (08 Marks) OR 2 a. Explain the following with respect to designing a leaming system : i) Choosing the training experience ii) Choosing the target function )) Choosing a representation for the target function (09 Marks) b. Write Find-S algorithm. Apply the Find-S for Table QU(c) to find maximally specific hypothesis. (06 Marks) ¢. Explain the concept of inductive bias. (05 Marks) Module-2 3 a. Explain the concept of decision tree learning. Discuss the necessary measures required to select the attributed for building a decision tree using ID3 algorithm. (U1 Marks) b. Explain the following with respect to decision tree leaning : i) Incorporating continuous valued attributes ii) Alternative measures for selecting attributes iii) Handling training examples with missing attribute values. (09 Marks) OR 4 a. Construct decision tree using ID3 considering the following training examples Weekend | Weather | Parental availability | Wealthy | Decision class Hi ‘Sunny Yes Rich Cinema He Sunn: No Rich Tennis [Hs [Windy] Yes_ Rich Cinema | He Rainy Yes Poor Cinema, Hs Rainy No Rich Home He Rainy ‘Yes Poor Cinema Th Wind: No Poor Cinema, He Windy No Rich Shopping Hy Windy Yes Rich Cinema, Ho Sunn; No Rich Tennis Table Q4(by (12 Marks) b. Discuss the issues of avoiding overfitting the data, and handling attributes with differing costs. (08 Marks) 1of217CS73 Module-3 5a. Discuss the application of neural network which is used to steer an autonomous vehicle. (06 Marks) b. Write Gradient descent algorithm to train a linear unit along with the derivation. (08 Marks) c. Discuss the issues of convergence, local minima and generalization, overfitting and stopping criterion. (06 Marks) OR 6 a. List the appropriate problems for neural network learning. (05 Marks) b. Define precepiron and discuss its training rule. (05 Marks) c. Show the derivation of back propagation training rule for output unit weights. (10 Marks) Module-4 7 a. Explain Bayes theorem and mention the features of Bayesian leaming. (07 Marks) b. Prove that a maximum likelihood hypotheses cab be used to predict probabilities. (08 Marks) c. Explain Naive Bayes classifier. (05 Marks) oR & a. Describe MAP leaming algorithm. (08 Marks) . b. Classify the test data and (Red, SUV, Domestic} using Naine Bayes classifier for the dataset shown in Table Q8(b). Color_[ Type | Origin Stolen Red _| Sports | Domestic | Yes Red_| Sports | Domestic | No Red_| Sports | Domestic | Yes Yellow | Sports | Domestic | No Yellow | Sports | Imported | Yes Yellow | SUV _| Imported | No | Yellow | SUV | Imported | Yes_ Yellow | SUV _| Domestic | No Red | SUV _| Imported | No Red_| Sports | Imported | Yes Table Q8(b) (06 Marks) €. Write and explain EM algorithm, (06 Marks) Module-5 9 a. Define i) Sample error True error ) Confidence intervals. (06 Marks) b. Explain K-nearest neighbor leaming algorithm. (08 Marks) ¢. Write a note no Q— leaning. (05 Marks) oR 10 a. Define mean value, variance, standard deviation and estimation bias of a random variable. (04 Marks) b. Explain locally weighted linear regression and radial basis functions. (10 Marks) ¢. What is reinforcement learning? How it differs from other function approximation tasks? (06 Marks) teeee 2of2 aaa5 9 7 fpgsh Proavam let Empwverts performance Mithine leamni Subject Code: (7 CS 7. at Rome fatk trrvovgyh Exper once, NeHAS any five + Lemming tev etegrine Raper] Wworeld W) Leaning te otsive on auteromows Veh Gi) Learning te Cladteby ree adbemomitol Ahucter oy) Leomming pey vortd- Cpls beuagammry W) Predactiney verovely nates of pneumonia pot Kyi) Detecting frouclulent use of Credil Carts Concept Leeming — Prabem % a predefimed Rpoce prteatiol \ypotketes ty kee Voqpetnetns tod best his We hesivy Cxorrpler Vox Torget Conmpt - aka — ol Leorre Task ~ PY Repreterlation ot Vypolter’ Oe &24<640,0, o>) Go + $4372, 2773 g,: {ART RR Y>Y Gy + $<2,7,2,%/2>} Sy: $<) a, Pe 4h X26 OS, 8,6, 974 Ga + 460.9,2, 1, 27KUV GY SLUR? 2 x,22R TRIP St 4207, & 27, BLU, PP?Pe | —— learniney Subject Code: C$ 7% ion . . Soluti mir Number on Allocated Sa. JSUT R2993 Xa7 CST, 2.4 Yt Ga Xo 9 “ $82 27-9077.7, 978.2. S17, P25} 7 x Sur [ORT RY xy LAT RG - Gu $42.7 Daher, DLP y rh | Set MALAY FR XerKESK YE Ges {Cu 7,4 y><72R yr} x vein pre Sinte Le R bs mekult oe ko to tnprelet s ore Same, prevind inte Ghintiance ix eve, We wht fate Grow fee Oy LOLDET2 2) J \ \ a BLUM LET BUI SVT YD \ ra “ | g: [Ru & we ¥rd BI chee ming be Treinrey SPENT Spat emer Cubs be Mauerte fast freetowr Dwet qpeqree to wath |e rho freedom, Rew-hree, < — pls bution baring Sromples ek BAH 4 1 Creassing Foagel fas tre Chaenve LBM Ateenale ve. RR {20 Represevtabon OH. wots PS a We We tts evne leconi Subject Code: 176 7% Solution Marks Allocated again: : ee tea weer Apertic hypothe Ht 2s fev Bah +e bveinivey Prttne . 2 For Gch attribute Contheint At \ Th tux Contbaint a Redioted by M Then clo nothing p wh by Clee repre @; 6 Saat fear > Zodt tect by ™ 3 extras hypotheses kez 40,0, 6, 0, o> we cRiT RR YT, bg > 22, T,2,9,97 hu 14T, RL Yyy be <2, 2,8 DD A Grred Hypitherry Sete —2 ; An Unbiedcol keavne® ae stan Fablity | bia fre leaning «| Concept & cleclioo. fico leony] - & Entropy (0) = Folge - war? 3 tabopy (2 SF} leg “ (sa exopy(S) Gon, Ad = cer (DZ fel _ 2 Cosatasse (8, A) 2 Cobnteia) Spit ota (¢,4) Dante mi Ue, Veda fos WO Te genera] hp =2R 7, RP yD lay S74 Vbiy pote Skoupled —Volxd otis ~B— Q prety Qudiaucrds= Vor oe “y 5 |S) Se i = Zag \Si Oy Atternadive wrankins Spline ad= 72 astaa Subject Code : 17 573 Marks Allocated Solution EUS)=N ST retest Goins, weatters=[0 6757) Grote (C., Parenti) > Goin ($) Wealtly)= 09246 pore Compr Pac Mie 3139 Hud Coin (Sssnny 1 Rewents) = (0°58 — fighew- Code (SSommq, Wolly> O [went | im (Spsnely, Poe) 2 OSB TAT Natty Da (Cyndy, weollingy= © 126 se oa f yom Coan (Geainy , Pareds) = oO att] ya / he Cece CCpomy, women) > 2°98} Chaema Tonal pele nT Civene Shbping 5) | aveating everyting data - Herding, tHe wll + Ly BM dilgeing Costs co Representa me — Explernieers —UY S | 3 treoaend-Bercot Ctmtn9 Srosgles, 9) oO ~~ 7) e) VTaphebve Gok wi je Rene Jawad Mawadon Vole L Dutt kaminction Quick OS = eg initeline enc Ass 400 L «Pex Sach (50,0) trot d Srouper, Pe a + Tagut Wx intetan te PB te we ont ecole lo + Per tach Qineay unt werdd ds, Do AWje Avwir n (to) Bncay Unt VEYs WH O2 wore Wt AW k petivetya~—- 4M | 64 © pot EachSubject Title: ine \eannivg Subject Code : | 7. Cg 7 Somber eat Atated 5@ Comarqente. & hoeel minima 12 Cenciertiration, Ovevdittinn, @ &b(P iy Crifettoy — 6a sy — SM D] Peceptar disgiem, Syrabirn, Srplomrobrs — ny Rainivg ule —]24 9D ae) 9 So) 9 19 Rous prapyphnn Rules Avis Sa , &a Ae ite. 3 ‘ \o Be Rolly Avs 2 SEE oy (to \U-GP “| Beges “The sre — 3% ¢ Frodunds ate a Prod? hy ery Zai tyhCas) + Cat) h(ihOu)) | 8m eA Explomalion =—| 4 ba 1. potas PCP) POD aa Ne O TPO Olkautin | BY a. mae > oro P&I) leu Pres = Ho 2Ok, Wier > 0 Pheedy Py) PESHToay” OCErre nA) = 01S 6 fre Pray PRAMAD PENA): POM 4) = 0 SK eye Gx o: 6 6m = BoD (Qed, SOV, OmeshO) — Yo romatine, = Ategnten —3, Sxplomediar - @ Hara ay 2X STE B Meprithn + Seplamasnn “Hy - B to 0 P : ais FD G-learning Gpmahr = |e Ay 2-34 DEReAXS 10M 0) abehu—