Reg. No. : | I Question Paper Code : 30148 M.C.A. DEGREE EXAMINATIONS, NOVEMBER/DECEMBER 2022, “Third Semester MC 4301 - MACHINE LEARNING (Regulations — 2021) Time : Three hours ‘Maximum : 100 marks Answer ALL questions, PART A — (10 x 2= 20 marks) 1. List out four types of data with appropriate example. 2. Mention the importance of data remediation. 3. Why is cross-validation good for testing? 4. What is meant by feature transformation? 5. How is maximum likelihood estimated? 6. What is the usage of EM algorithm? 7. Compare and contrast on: biological neuron and Artificial Neuron. % Highlight the role played by Bias and Variance in ML. 9. Compare and contrast on: discrete and Continuous Attribute. 10. Why Hinge loss function is important in SVM? PART B— (6 x 13 = 65 marks) 11. (a) Explain the steps involved in data preprocessing with an example. Or (b) Explain the steps involved in Machine learning model preparation.12. 13. 14, 15, 16. (a) @) @) ) @ ) (a) b) (a) (b) Elaborate on the feature selection methods to reduce dimensionality, Or How do you evaluate and improve the performance of a model? Explain with suitable example. Explain about components of Bayesian Belief networks with an example. Or Explain the process of maximum likehood and least square error in bayesian learnign. Ilustrate the different type’s logistic regression with suitable example. Or Illustrate the significance of the activation functions. Tanh, Sigmoid, Relu and PRelu. Examine how decision trees are effective as classifier. Or Illustrate about the SVM kernels with suitable examples. PART C—(1x 15 = 15 marks) Expand on Bayes optimal classifier and apply in to the following scenario: Vincent has a lot of folders in office to classify as business or family. He has classified the first bookshelf full of folders and is using this to estimate the probabilities that business or family folders have particular words in their lables. P(Travel/Family)=0.4; (‘Travel/Business)=0: P(Records'Family)=0.3; P(Records'/Business)=0.6; P(Spain/Family)=0.2; P(Spain/Business)=0.1;_P(Family)=0.4; P(Business)=0.6; Using these estimates, he decides to classify future folders. What estimate will Vincent get for P(Pamily/'Spain’, ‘Travel’, ‘Records? Or Evaluate the Perceptron Neural Network in detail. Also using Perceptron find the new weights: Vectors: (-1, 1, -1, -I), (1,1y-1,-D) belong to class 1(1) and Vectors: (-1, 1, 1, -1), (1, -1, 1, 1) belong to class 2(-1). Use a learning rate of 1 and starting weights of 0. Using the training vectors as input, test the response of the net using algorithm. 2 30148Reg. No. : Question Paper Code : 30149 M.C.A. DEGREE EXAMINATIONS, NOVEMBER/DECEMBER 2022. ‘Third Semester ‘MC 4302 ~ INTERNET OF THINGS (Regulations 2021) ‘Time : Three hours Maximum : 100 marks 10. ll. Answer ALL questions. PART A— (10 x 2= 20 marks) What is the role of controller service in an Tot system? What is the difference between embedded sys and IoT system? List few protocols used in M2M. Mention any two major issues relevant to ToT standardization, What is Raspbian OS? What is Broadcom Numbering (BCM) of Raspberry Pi? Why is data aggregation needed for Io applications? What are FP7 projects? List some features on smart manufacturing. What is a middleware in Tot? PART B— (6 x 13 = 65 marks) (a) List the two communication APIs of Io. Explain the them in detail. Or () Describe an example of IoT service that uses publish subscribe communication model. 512. 13, 14. 15. 16. (a) (a) ) @) ©) @) (b) @ (b) Explain the communication protocols that are used for M2M local area networks in detail. Or What are the purposes of BACNet protocol and KNX. Explain in detail. () Explain various interfaces available in Raspberry Pi with appropriate examples. a) (i) Write a simple Arduino program to blink a LED with 500 msec, delay in between ‘ON’ and OFF’. © Or () What are the differences in hardware and software serial in Arduino? Explain each with appropriate application. oy (ii) Write a simple Arduino program to read the analog joystick value and display the direction of joystick on serial Monitor. ©) Discuss the importance of the terminologies ‘Trust for IoT’ and ‘Privacy for lot’. Or Brief a methodology to obtain secured platforms for smart cities. And also explain the process of data aggregation for the Jot in smart cities. What are Iot Smart-X applications? List a few and explain any one in detail with a diagram. Or Explain the future factory concept of IoT in detail with appropriate mind map or flow diagram. PART C— (1x 15 = 15 marks) Determine the Io levels for designing home automation IoT systems including smart lighting and intrusion detection. Explain with neat diagram. Or Design a simple IoT application using Digital /O in RPi3 with suitable connection diagram. Write and explain a suitable python seript for the same. 2 30149 2b12, 13. 4. 16. 16. (a) (b) (a) (b) (a) &) (a) (b) (a) (b) Explain the communication protocols that are used for M2M local area networks in detail. Or What are the purposes of BACNet protocol and KNX. Explain in detail. @ Explain various interfaces available in Raspberry Pi with appropriate examples. @ (i) Write a simple Arduino program to blink a LED with 500 msec, delay in between ‘ON’ and OFF’ © Or (@ What are the differences in hardware and software serial in Arduino? Explain each with appropriate application. mM (ii) Write a simple Arduino program to read the analog joystick value and display the direction of joystick on serial Monitor. © Discuss the importance of the terminologies “Trust for Jol” and ‘Privacy for lot’. Or Brief a methodology to obtain secured platforms for smart cities. And also explain the process of data aggregation for the Iot in smart cities, What are Iot Smart-X applications? List a few and explain any one in detail with a diagram. Or Explain the future factory concept of Io in detail with appropriate mind map or flow diagram. PART C — (1x 15 = 15 marks) Determine the IoT levels for designing home automation IoT systems including smart lighting and intrusion detection. Explain with neat diagram. Or Design a simple IoT application using Digital I/O in RPi3 with suitable connection diagram. Write and explain a suitable python script for the same. 2 30149see mes TET ELI Question Paper Code : 30152 M.C.A, DEGREE EXAMINATIONS, NOVEMBER/DECEMBER 2022, Third Semester (Bridge Course) BX 4006 — BASICS OF COMPUTER NETWORKS (Regulations 2021) ‘Time : Three hours Maximum : 100 marks 3, ex 10. u, Answer ALL questions, PART A — (10 x 2= 20 marks) What is encapsulation? How many links are needed to connect five devices using mesh, star, bus and ring topology network? List some of the error detection techniques. Define bit stuffing. Mention the use ICMP protocol. An organization is granted the block 221.17.180.0/24. The administrator wants to create 32 subnets. Find the subnet mask. Compare cireuit switching and packet switching. List some of the routing protocols. Mention the use of DNS server. Compare threat and vulnerability. PART B— (6 x 23 = 65 marks) (a) With a neat sketch, explain the role played by the seven layers in the OSI ‘model to make communication possible among heterogeneous systems, Or (b) How guided media differ from unguided media? Briefly explain any three mothods used for data transmission using guided media and two methods used for data transmission using unguided media, RT12, 13, (@) ) (a) ) () Draw the process flowchart for CSMA/CA used in wireless LAN and explain it in detail. @ (i) Draw the Ethernet Frame Format and explain the purpose of each field. © Or Obtain the Frame Check Sequence generated by the CRC generator that, uses polynomial divisor X++X+1 on the data stream of 1101011011 and show how the reesiver checks the received bit for no error. ‘An ISP is granted a block of addresses starting with 160.40.0.0/16. The ISP wants to distribute these blocks to 3 groups of customers as follows. (Group I has 64 customers, each need 256 addresses. (ii) Group TT has 128 customers, each need 128 addresses (iii) Group III has 128 customers, each need 64 addresses. Design the sub-blocks and give the slash notation for cach sub- block. Find out how many addresses are still available after these allocations, Or Explain the setup, data transfer and teardown process in data transmission of any network. Below table shows the types of networks and the addressing mechanism used in each of them. Network Setup | Data transfer | Teardown Cireuit-Switched | End-to-End = End-to-End Datagram - = Virtual-circuit | End-to-End End-to-End Answer the following questions () Why does a circuit-switched network need end-to-end addressing during the setup and teardown phases? Why are no addresses needed during the data transfer phase for this type of network? (5) (ii) Why does a datagram network need only end-to-end addressing during the data transfer phase, but no addressing during the setup and teardown phases? @ (iii), Why does a virtual-cireuit network need addresses during all three phases? ® 2 30152 a14, 16. () (b) (a) (a), Draw a UDP Header format and explain each one of the field in it. The following is a dump of a UDP header in hexadecimal format. (CB84000D001C001C),, (What is the source port number? Gi) What is the destination port number? Gi) What is the total length of the user datagram? (iv) What is the length of the data? Or Explain in detail the connection establishment, data transfer and termination procedure of TCP using three-way handshake protocol with an example, Dosoribe the general architecture and working principle of an SMTP in Internet for sending an e-mail. Or Write down the algorithm of RSA. In a public key crypto system using RSA algorithm the intruder intercepts the cipher text as 11 and whose public keys are (7,187). Find the private key value d and perform encryption and decryption process. PART C— (1x 15= 15 marks) Draw the flowchart of the link state routing algorithm. Compute the shortest path for the node A in figure below using Dijkstra’s path algorithm. For each step in the algorithm, indicate the current node, the assigned labels and whether the labels are permanent or tentative. Finally, list the shortest path to all destination nodes as seen from the node A. 8 30152(b) Draw a neat sketch of TCP header format and explain each field in it. Work out the following is a TCP header in hexadecimal from (05320017 00000001 00000001 500207 FF 00000000), To find, () Source port number (i) Destination port number (iii) Sequence number (iv) Acknowledgement Number (v) Length of the header (vi) Type of the segment (vii) Window Size. 4 30152LT J Question Paper Code : 60881 M.C.A. (2 years) DEGREE EXAMINATIONS, NOVEMBER/DECEMBER 2022. Reg. No.: E fl ‘Third Semester MC 5306 — DATA SCIENCE (Regulations 2017) ‘Time : Three hours ‘Maximum ; 100 marks 10. Answer ALL questions. PART A — (10 x 2= 20 marks) Paraphrase about Monte Carlo Resampling technique. Additionally, outline why a Monte Carlo simulation approach focuses on constantly repeating random samples. In a time interval of 10-minutes. The probability that a person may see a shooting star or a bunch of them is 0.3, What is the percentage chance of that person seeing at least one star shooting from the sky if the person is under it for about an hour? List down the conditions for Overfitting and Underfitting With an assumption of a null hypothesis as correct, what docs it mean when the p-values are high and iow? Define Root Mean Square Error (RMSE) and Mean Square Hrror (MSE) in a linear regression model. Define the term Kurtosis, Give the formula for the same. What does a positive Fishers’ Kurtosis value mean to us? List down the cardinal weakness of a schema-based database, What is schema- less database? Give examples. List down its pros and cons, Identify the reason behind the shuffling of partitions of data in a MapReduce (MR) framework. Why is it mandatory to filter a data stream? List down any two filtering techniques pertaining to data mining, Outline the approach of Tomek links specific to under sampling. By removing the Tomek links from the dataset, would we expect the resulting transformed dataset to be balanced? Justify your point. 10 3|u 12, @ ) (@) & @ @ Gi) @ wa «a a Gi) «i PART B — (6 x 13 = 65 marks) Frame an approach for solving any data analytics based project. (6) How do you identify if a coin ¢ is biased or not? What null hypothesis claims with respective to the coin ‘c’? oO Or Give short notes on Conditional Inference Trees. © List down any two light-weight but powerful web data visualization tools. In what way they are efficient/powerful? Explain in brief about the mentioned tools. M Specification: Toss the solected coin 10 times from a jar of 1000 coins. Out of 1000 coins, 999 coins are fair and 1 coin is double-headed, assume that you see 10 heads. Estimate the probability of getting a head in the next coin toss. (6) List down the various assumptions used in linear regression? What would happen if they are violated? Give suitable examples to justify your point, ” Or Discern the differences between correlation and covariance. ©) Tabulate the differences between univariate, bivariate and multivariate Analysis with an example. a Outline the Philosophy behind CAP Theorem. 6) Paraphrase about Support Vector Machine (SVM)? Is it good to do dimensionality reduction before fitting a Support Vector Model? Describe how a kernel function serves the purpose of fitting an optimal hyper plane while classifying data available in a k-dimensional plane, ®) Or Outline the methodology behind a kernel trick and enunciate about the noed of selection bias. © Assume that in a spring boot application that receives data from a user ‘u', the data contain few images and videos. How these entities (big files) are stored in database(MongoDb) provided it is not possible to access Amazon $3. Will the docker on a VM be useful for this scenario? If so, in what way it is useful? @ 2 60881 3a14, 16 @) (b) fa) ) (a) @ (i) Gi) ‘How the MapReduce sort algorithm works in general? With a simple input instance, show how the MapReduce sort algorithm works and specify the ordering of arrangement of numbers in each phase of the algorithm, © List down the key differences to do Map/Reduce (MR) work on MongoDB using Hadoop-MR vs built-in MR of Mongo? Outiino the circumstances/sconarios to pick which MR engine works effectively. Outline the pros and cons of each engine to work on data stored in mongodb? a) Or Brief about the various components of Hadoop Distributed File System (HDFS), Explain the functions of HDFS in regard to monitoring, maintaining, integrity and replication. © ‘The default block size of HDFS is 64 MB/128 MB while the default, block size of UINX/LINUX is 4 KB/128 KB. Why? What implication do you think, this will have in the design of the NameNode? (7) Using DGIM (Datar-Gionis-Indyk-Motwani) Algorithm, show how the bits that are flowing through a sliding window can be counted. Prove that the storage area consumes O(logs/V) space compiexity with N - number of queries within a stipulated time interval [f, &]. Or Devise the Flajolet-Martin algorithm (in pseudocode) that estimates the number of distinct elements, With a suitable input instance, demonstrate hovw the estimates of distinet elements are done in each and every phase of the prescribed algorithm, Does the algorithm run in polynomial time(Py,0)? Justify. ws PART C— (Lx 15 = 16 marks) Given a data stream consists of the positive iatogers 3, 1, 4, 1, 5, 9, 2, 6, 5 and the adopted hash functions are of the form h(x) = ax+ b mod 82 for some a and b. The result, is treated as a 5-bit binary integer. Determine the tail length for each stream element and the resulting estimate of the number of distinct elements if the hash function is: @) (@) h(@)=2x+1mod32 @ A(x) =3x+7mod 32 @) h(x)=4xmods2, Describe how oyer sampling mitigated using SMOTH (Synthetic Minority OverSampling Technique). o Or 3 60881 33® @ Gi) ‘Assume that we have a stream of items of large and unknown length that we can only iterate over once. Devise an effective sampling algorithm that randomly chooses an item from this stream such that each item is equally likely to be selected. @ Assume that you have n-bits of memory available and the set of key values $ has m members. While using Bloom's filter, instead of possessing k hash functions, divide the n-bits into k arrays and hash one to each array, find the probability of a False-Positive as a function of n, m and h. om 4 60881