Scribd
Upload
124 views7 pages

Database and Machine Learning Exam Papers

The document outlines examination papers for various subjects including Database Management System, Machine Learning Applications, Deep Learning, Soft Computing, Data Mining, and Computer Networks, each with a structured format of questions categorized into Very Short Answer, Short Answer, and Long Answer types. Each paper has a total of 70 marks and requires candidates to answer questions in their own words. The questions cover fundamental concepts, applications, and technical details relevant to each subject area.

Uploaded by

kunduakash1102
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
124 views7 pages

Database and Machine Learning Exam Papers

The document outlines examination papers for various subjects including Database Management System, Machine Learning Applications, Deep Learning, Soft Computing, Data Mining, and Computer Networks, each with a structured format of questions categorized into Very Short Answer, Short Answer, and Long Answer types. Each paper has a total of 70 marks and requires candidates to answer questions in their own words. The questions cover fundamental concepts, applications, and technical details relevant to each subject area.

Uploaded by

kunduakash1102
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

CS/B.

TECH(N)/EVEN/SEM-6/6910/2023-2024/I019
MAULANA ABUL KALAM AZAD UNIVERSITY OF TECHNOLOGY, WEST BENGAL
Paper Code : OECAIML 601A Database Management System
UPID : 006910

Time Allotted : 3 Hours Full Marks :70


The Figures in the margin indicate full marks.
Candidate are required to give their answers in their own words as far as practicable

Group-A (Very Short Answer Type Question)


1. Answer any ten of the following : [ 1 x 10 = 10 ]
(I) What are the different schema in dbms?
(II) What are the different types of constrains in DBMS?
(III) what is orderby clause?
(IV) Explain two types of query language.
(V) Explain closure property with example?
(VI) Name all the data models in DBMS.
(VII) What is Serializability?
(VIII) What is DDL in DBMS?
(IX) what is SQL?
(X) what is BCNF?
(XI) what is DML in DBMS?
(XII) what is group by clause ?

Group-B (Short Answer Type Question)


Answer any three of the following : [ 5 x 3 = 15 ]
2. Explain mapping cardinality using ER diagram. [5]
3. Explain Specialization with suitable example. [5]
4. Explain views in DBMS. [5]
5. What are the different joins in DBMS? [5]
6. Explain Armstrong Axioms with suitable examples. [5]

Group-C (Long Answer Type Question)


Answer any three of the following : [ 15 x 3 = 45 ]
7. Draw an ER diagram for Library Management System. How you can converrt ER model to Relational [ 15 ]
model.
8. Describe with suitable example in relational algebra [ 15 ]
i) Union ii) Natural Join iii) Intersection iv) Set Difference v) Selection and Projection
9. If a relation R(ABCDEF) and following are the functional dependencies. [ 15 ]
AB->C
C->DE
E->F
F->A
Check the highest normal form.
Explain the Multivalued Dependency(MVD) with suitable example.
10. Explain different types of keys in DBMS with examples . Explain different types of indexing in RDBMS. [ 15 ]
11. What is transaction? Explain ACID properties of transaction. A relation R(A,B,C,D,E) and FDs are A- [ 15 ]
>BC,CD->E,B->D,E->A find all the candidate keys?

*** END OF PAPER ***

1/1
CS/[Link](N)/EVEN/SEM-6/6903/2023-2024/I019
MAULANA ABUL KALAM AZAD UNIVERSITY OF TECHNOLOGY, WEST BENGAL
Paper Code : PCCAIML601 Machine Learning Applications
UPID : 006903

Time Allotted : 3 Hours Full Marks :70


The Figures in the margin indicate full marks.
Candidate are required to give their answers in their own words as far as practicable

Group-A (Very Short Answer Type Question)


1. Answer any ten of the following : [ 1 x 10 = 10 ]
(I) ___________ is the clustering method which takes care of variance in data.
(II) ______________ clustering technique requires a merging approach.
(III) __________ K stand for in K mean algorithm?
(IV) __________clusters formed in this method forms a tree-type structure based on the hierarchy.
(V) Concept learning is also known as ____________
(VI) The Candidate-Elimination Algorithm represents the ____________.
(VII) Decision tree is a ____________ algorithm.
(VIII) Suppose, your target variable is the price of a house using Decision tree. What type of tree do you need to
predict the target variable?
(IX) _________ algorithm is used for solving temporal probabilistic reasoning.
(X) __________ is the machine learning algorithms that can be used with unlabeled data.
(XI) __________ is not a supervised learning.
(XII) Deep reinforcement learning has produced _____________

Group-B (Short Answer Type Question)


Answer any three of the following : [ 5 x 3 = 15 ]
2. Compare SVM and Logistic Regression in handling outliers. [5]
3. What are the differences between Linear Regression and Logistic Regression? [5]
4. What do you understand by Reinforcement Learning technique? [5]
5. What are some applications of Unsupervised Learning? [5]
6. How can overfitting be avoided? [5]

Group-C (Long Answer Type Question)


Answer any three of the following : [ 15 x 3 = 45 ]
7. (a) What is ‘Training set’ and ‘Test set’? [5]
(b) How do classification and regression differ? [5]
(c) What do you understand by Accuracy? [5]
8. (a) Which should be preferred among Gini impurity and Entropy? Explain. [7]
(b) What do you understand about Information Gain? Also, explain the mathematical formulation [8]
associated with it.
9. (a) List down some popular algorithms used for deriving Decision Trees and their attribute selection [5]
measures.
(b) Explain the CART Algorithm for Decision Trees. [5]
(c) Briefly explain the properties of Gini Impurity. [5]
10. (a) What do you understand by the Confusion Matrix? [7]
(b) Explain True Positive, True Negative, False Positive, and False Negative in Confusion Matrix with an [8]
example.
11. (a) What is Bagging and Boosting? [5]
(b) What do you understand by Cluster Sampling? [5]
(c) Why are PCA, KPCA and ICA used? [5]

1/1
CS/[Link](N)/EVEN/SEM-6/6904/2023-2024/I019
MAULANA ABUL KALAM AZAD UNIVERSITY OF TECHNOLOGY, WEST BENGAL
Paper Code : PCCAIML602 Deep Learning
UPID : 006904

Time Allotted : 3 Hours Full Marks :70


The Figures in the margin indicate full marks.
Candidate are required to give their answers in their own words as far as practicable

Group-A (Very Short Answer Type Question)


1. Answer any ten of the following : [ 1 x 10 = 10 ]
(I) Neural Networks are complex ______________ with many parameters.
(II) What is "Rectified Linear Unit Function"?
(III) Bayes network consist of ....
(IV) Why linearly separable problems are of interest for neural network researchers?
(V) Define Entropy.
(VI) A 4-input neuron has weights 1, 2, 3 and 4. The transfer function is linear with the constant of proportionality
being equal to 2. The inputs are 4, 10, 5 and20 respectively. What will be the output?
(VII) What is sigmoid function ?
(VIII) In a Convolutional Neural Network (CNN), how can you fix the constant validation accuracy?
(IX) What is Batch Gradient Descent?
(X) What do you mean by an epoch in the context of deep learning?
(XI) What are the advantages of transfer learning?
(XII) What is Sparse Coding?

Group-B (Short Answer Type Question)


Answer any three of the following : [ 5 x 3 = 15 ]
2. What is the meaning of overfitting? [5]
3. Explain the application of deep learning in Object Recognition. [5]
4. Explain Hidden Markov Model. [5]
5. What is the use of the loss function? [5]
6. Explain backpropagation in brief. [5]

Group-C (Long Answer Type Question)


Answer any three of the following : [ 15 x 3 = 45 ]
7. (a) What are the different types of supervised learning? Explain with example. [ 10 ]
(b) State the differences between Supervised and Unsupervised learning. [5]
8. (a) What is "Log loss"? [5]
(b) Explain the concept of penalty and reward in reinforcement learning? [ 10 ]
9. Explain the architecture of convolutional neural network (CNN) with suitable diagram. [ 15 ]
10. (a) State the differences between Deep and Shallow Network. [5]
(b) Explain Recurrent Neural Network (RNN) and its applications in Deep Learning. [ 10 ]
11. Explain PCA (Principal Component Analysis) and its applications in brief. [ 15 ]

*** END OF PAPER ***

1/1
CS/[Link](N)/EVEN/SEM-6/6905/2023-2024/I019
MAULANA ABUL KALAM AZAD UNIVERSITY OF TECHNOLOGY, WEST BENGAL
Paper Code : PCCAIML603 Soft Computing
UPID : 006905

Time Allotted : 3 Hours Full Marks :70


The Figures in the margin indicate full marks.
Candidate are required to give their answers in their own words as far as practicable

Group-A (Very Short Answer Type Question)


1. Answer any ten of the following : [ 1 x 10 = 10 ]
(I) Computational power in ANNs is determined by _______.
(II) The essential sites for recombination are known as ______.
(III) How many types of random variables are there in fuzzy logic?
(IV) What is the main intention of the defuzzyfication height method?
(V) Write some applications of Genetic Algorithm.
(VI) Which process is carried out to separate chromosomal or genomic DNA?
(VII) Automated vehicle is an example of ______.
(VIII) Ultimate results are produced in ANN at ________.
(IX) Error correction learning is like "learning with teacher". Is it true or false?
(X) What is Simulated Annealing?
(XI) Write down one example of active learning.
(XII) In membership function graph, X axis represents which parameter?

Group-B (Short Answer Type Question)


Answer any three of the following : [ 5 x 3 = 15 ]
2. Write down the difference between hard computing and soft computing. [5]
3. Explain how fuzzy logic is applied in home appliances? [5]
4. State the difference between Adaline and Madaline networks. [5]
5. Explain Crossover and Mutation with examples. [5]
6. Write down the difference between crisp logic and fuzzy logic. [5]

Group-C (Long Answer Type Question)


Answer any three of the following : [ 15 x 3 = 45 ]
7. Explain GA based clustering algorithm. [ 15 ]
8. (a) The elements in two sets A and B are given as A = {2, 4} and B = {a, b, c}. Find the various Cartesian [ 10 ]
product of these two sets.
(b) Explain convex fuzzy set and non-convex fuzzy set. [5]
9. Write Short Notes on Particle Swarm Optimization (PSO). [ 15 ]
10. Explain Mamdani fuzzy system. [ 15 ]
11. Design a Hebb net to implement logical AND function. [ 15 ]

*** END OF PAPER ***

1/1
CS/[Link](N)/EVEN/SEM-6/6907/2023-2024/I019
MAULANA ABUL KALAM AZAD UNIVERSITY OF TECHNOLOGY, WEST BENGAL
Paper Code : PECAIML 601B Data Mining
UPID : 006907

Time Allotted : 3 Hours Full Marks :70


The Figures in the margin indicate full marks.
Candidate are required to give their answers in their own words as far as practicable

Group-A (Very Short Answer Type Question)


1. Answer any ten of the following : [ 1 x 10 = 10 ]
(I) Define data warehouse.
(II) Define underfitting.
(III) What is testing data?
(IV) Provide one real-life example of regression analysis.
(V) Write down the full form of LARS.
(VI) What is lift curve?
(VII) Define sparse data.
(VIII) Define noise.
(IX) Define standardization of data.
(X) Define the concept of high bias in relation training a model.
(XI) Mention any important feature of random forest.
(XII) Mention what learning rate defines about network parameters.

Group-B (Short Answer Type Question)


Answer any three of the following : [ 5 x 3 = 15 ]
2. Write down the concept of linear regression with example. [5]
3. State and explain the major tasks involved in data preprocessing. [5]
4. Write down and explain the principle of Naive Bayes classifier. [5]
5. State and discuss the major steps involved in the task of knowledge discovery. [5]
6. Explain with example the geometric intuition behind SVM? [5]

Group-C (Long Answer Type Question)


Answer any three of the following : [ 15 x 3 = 45 ]
7. (a) Discuss the various major components of a mining system with proper diagram. [ 10 ]
(b) Write down any 5 differences between prediction and classification. [5]
8. (a) Explain feed forward neural network with respective relations and diagrams. [8]
(b) Write down the perceptron model with proper representation and diagram. [7]
9. (a) Explain Decision tree classifier algorithm with proper diagram and mention the algorithm. [ 10 ]
(b) Explain any one (1) of the following terminologies associated with decision tree: [5]
Information Gain
OR
Gini Index
10. (a) What is association rule mining? Mention the steps associated it. [3]
(b) Mention the various metrics associated with association rule mining and describe them briefly with [9]
respective mathematical expressions.
(c) What is F-P tree? Write any two advantages of it. [3]
11. (a) Define confusion matrix and explain the same with corresponding metric name for each cell. [6]
(b) Define the following terms derived from confusion matrix with related equations. [9]
Sensitivity
Specificity
Precision

1/1
CS/[Link](N)/EVEN/SEM-6/6596/2023-2024/I019
MAULANA ABUL KALAM AZAD UNIVERSITY OF TECHNOLOGY, WEST BENGAL
Paper Code : EC602/PCC-CS602/PCCCS602 Computer Networks
UPID : 006596

Time Allotted : 3 Hours Full Marks :70


The Figures in the margin indicate full marks.
Candidate are required to give their answers in their own words as far as practicable

Group-A (Very Short Answer Type Question)


1. Answer any ten of the following : [ 1 x 10 = 10 ]
(I) What is DNS in computer network?
(II) Which multiplexing technique shifts each signal to a different carrier frequency?
(III) What is the Hamming Distance between two equal codewords?
(IV) What are autonomous systems in computer networks?
(V) TCP assigns a sequence number to each segment that is being sent. The sequence number for each segment is
the number of the _______ byte carried in that segment.
(VI) What are the transmission modes in FTP?
(VII) In synchronous TDM, for n signal sources of the same data rate, what will be the number of slots in each frame?
(VIII) In bit stuffing when is an extra 0 bit stuffed to the data section of the frame?
(IX) What are the number of bits required for IPv4 addresses when expressed in binary?
(X) In the ____ bucket algorithm, bursty chunks are stored in bucket and sent out at an average rate.
(XI) Why is PGP needed?
(XII) The point where the secure internal network and untrusted external network meet and a firewall is installed is known as______ .

Group-B (Short Answer Type Question)


Answer any three of the following : [ 5 x 3 = 15 ]
2. How does FTP Work? [5]
3. Describe the functions of session layer of the OSI reference model. [5]
4. What is piggybacking? Discuss its advantages. [5]
5. What is the first address of a block of classless addresses if one of the addresses is [Link]/27 [5]
6. Suppose that the UDP receiver computes the Internet checksum for the received UDP segment and finds [5]
that it matches the value carried in the checksum field. Can the receiver be absolutely sure that no bit
errors have occurred? Explain. Would things be different with TCP?

Group-C (Long Answer Type Question)


Answer any three of the following : [ 15 x 3 = 45 ]
7. (a) What are the functions of MAC sublayer? [4]
(b) Discuss the working principle of pure ALOHA and slotted ALOHA. What is the vulnerable time in [ 3+2 ]
pure and slotted ALOHA?
(c) With diagram explain Go-Back-N protocol. [6]
8. (a) What do you mean by private IP address? Why is it needed? [5]
(b) Explain the DHCP state transition diagram. [6]
(c) A block of 16 addresses is granted to a small organization. If one of the addresses is [4]
[Link]/28, what are the first and last addresses in the block?
9. (a) With diagram explain three way handshaking in TCP. [6]
(b) What is the importance of port number? Consider a TCP connection between Host A and Host B. [ 2+1 ]
Suppose that the TCP segments travelling from Host A to Host B have source port number x and
destination port number y. What are the source and destination port numbers for the segments
travelling from Host B to Host A?
(c) Explain briefly how the leaky-bucket and token bucket algorithm works. How can these two [ 4+2 ]
algorithms (leaky -bucket and token-bucket) be combined to get the most optimized quality of
service?
1/2
10. (a) How does a packet filter firewall work? [7]
(b) With a block diagram explain symmetric key encryption. [8]
11. (a) Explain the two node instability problem in Distance vector routing. [5]
(b) Discuss the possible solutions of the two node instability problem in Distance vector routing. [5]
(c) Assume a packet has to visit two routers during transmission from source to destination. Determine [5]
how many times the packet has to visit the network layer and the data link layer.

*** END OF PAPER ***

2/2

Common questions

Powered by AI

ACID properties, which stand for Atomicity, Consistency, Isolation, and Durability, ensure data integrity in a database system by making sure transactions are processed reliably. Atomicity ensures that all parts of a transaction are completed successfully or none at all. Consistency ensures that a transaction brings the database from one valid state to another. Isolation ensures that the execution of transactions concurrently does not affect their outcome. Durability ensures that once a transaction has been committed, it will remain so, even in the event of a system failure .

Mapping cardinality in an ER diagram is significant because it determines the number of instances of one entity that can be associated with instances of another entity. It is represented in the ER diagram by placing annotations on the lines connecting different entities. These annotations specify the lower and upper bounds of the number of associations. For instance, a one-to-many relationship would be represented by '1' on one side and 'N' on the other .

Supervised learning differs from unsupervised learning primarily in terms of data labeling and objectives. In supervised learning, models are trained on labeled data, meaning each input comes with a corresponding output. The goal is to learn a mapping from inputs to outputs, making it ideal for tasks like classification and regression. Unsupervised learning, on the other hand, deals with unlabeled data, aiming to find hidden structures or patterns in the data, such as clustering or dimensionality reduction. This difference defines their use cases, data requirements, and outcomes in machine learning processes .

PCA enhances data analysis in machine learning by reducing the dimensionality of the data, which simplifies the dataset and enables better visualization and understanding of the data's structure. PCA transforms the data into a set of orthogonal components that capture the maximum variance, effectively reducing the number of variables while preserving as much information as possible. This process helps in eliminating redundant features, decreases computational cost, and reduces the risk of overfitting, thus improving the performance of the learning algorithms .

Fuzzy logic is applied in home appliances to improve user experience and functionality by mimicking human decision-making. Common applications include washing machines, which adjust water levels and washing cycles based on the load size and type, and air conditioners, which regulate temperature and fan speed based on environmental conditions and user preferences. The benefits of using fuzzy logic in these appliances include increased efficiency, customization, adaptation to varying inputs, and improved energy savings, as these devices can operate with a level of nuance similar to human reasoning .

The leaky-bucket algorithm manages network traffic by ensuring a constant output rate, smoothing bursty traffic loads. It functions by letting packets flow into a fixed-capacity bucket and leaking them at a steady rate to the network. The token-bucket algorithm, on the other hand, allows bursty traffic by permitting packets to transmit until a token limit is reached, accommodating peak flows while maintaining an average rate. Combining these algorithms provides a robust strategy for managing traffic. The leaky-bucket ensures network stability and pollution prevention, while the token-bucket allows flexibility and maximizes throughput, optimizing the quality of service .

SVM is typically more effective at handling outliers compared to Logistic Regression. SVM tries to find the optimal hyperplane that maximizes the margin between different classes, inherently making it robust to outliers, as outliers do not influence the boundary line much once it is beyond the support vectors. Logistic Regression, on the other hand, uses a probabilistic approach and is significantly affected by outliers, as these can skew the boundary line due to its reliance on a linear decision boundary in the input space .

Hard computing relies on precise and exact mathematical models and is suitable for problems with well-defined parameters. It uses deterministic algorithms that follow a strict structure. In contrast, soft computing is designed to mimic human reasoning and is well-suited for vague or uncertain problems. It uses techniques like fuzzy logic, genetic algorithms, and neural networks to adapt to changes and improve over time. While hard computing focuses on finding optimized and accurate solutions, soft computing emphasizes flexibility and robustness in dealing with real-world ambiguities .

Backpropagation in neural networks is a method used to calculate the gradient of the loss function with respect to all the weights in the network. It works by propagating the error from the output layer back to the input layer and adjusting the weights using gradient descent to minimize the error. This process is essential for training neural networks, as it allows the network to learn from errors and improve its accuracy. Backpropagation is significant in deep learning as it enables the efficient training of deep networks, facilitating the adjustment of parameters in networks with multiple hidden layers, thus powering complex models used in various applications .

A confusion matrix contributes to the evaluation of a machine learning model by providing a detailed breakdown of the model's performance on each class. It shows the true positives, true negatives, false positives, and false negatives, enabling the calculation of key metrics such as accuracy, precision, recall, and F1-score. These metrics give insights into the model’s strengths and weaknesses, helping improve model development by highlighting where predictions are going wrong and areas needing improvement or greater focus .

You might also like