Jawaharlal Nehru Technological University Hyderabad
Jawaharlal Nehru Technological University Hyderabad B.Tech IV MACHINE LEARNING Previous Question Paper (July - 2019 - R15)
Note: This question paper contains two parts A and B. Part A is compulsory which carries 25 marks. Answer all questions in Part A. Part B consists of 5 Units. Answer any one full question from each unit. Each question in Part B carries 10 marks.
Part A
What are the issues in machine learning?
Answer
The primary issues in machine learning include the bias-variance tradeoff, which balances model complexity and generalization, and overfitting, where a model captures noise instead of the underlying pattern. Additionally, challenges include data quality (noise, missing values), computational complexity, and the requirement for large, representative datasets.
How to choose a function approximation algorithm?
Answer
Choosing a function approximation algorithm depends on several factors: the nature of the target function (linear vs. non-linear), the amount of available training data, the presence of noise, and the requirement for model interpretability. One should evaluate the bias-variance characteristics, computational constraints, and the dimensionality of the input space to select the most suitable model.
What is preference bias?
Answer
Sign in to view the complete answer paper.
How can you avoid over fitting the data?
Answer
Sign in to view the complete answer paper.
Define Confidence Interval?
Answer
Sign in to view the complete answer paper.
What is conditional independence?
Answer
Sign in to view the complete answer paper.
What are the limitations of eager learning?
Answer
Sign in to view the complete answer paper.
Why do we need locally weighted regression?
Answer
Sign in to view the complete answer paper.
How can you reduce prior knowledge for reducing sample complexity?
Answer
Sign in to view the complete answer paper.
What are the ways of parallelizing genetic algorithms?
Answer
Sign in to view the complete answer paper.
Part B
Write about List-Then-Elimination algorithm.
Answer
Sign in to view the complete answer paper.
Define the following: i) Inductive Learning Hypothesis ii) Consistent hypothesis iii) Version Space for the PlayTennis problem.
Answer
Sign in to view the complete answer paper.
With an example, Explain the working of Find-S algorithm?
Answer
Sign in to view the complete answer paper.
Discuss in brief about Inductive bias.
Answer
Sign in to view the complete answer paper.
Explain about decision tree based learning? How it is represented. Give some problems for which decision tree based learning is appropriate.
Answer
Sign in to view the complete answer paper.
Explain ID3 algorithm for decision tree learning.
Answer
Sign in to view the complete answer paper.
Define neural network learning. What are the problems in neural network learning?
Answer
Sign in to view the complete answer paper.
Explain in brief about Back propagation algorithm.
Answer
Sign in to view the complete answer paper.
Define sampling theory. What are the basics of sampling theory?
Answer
Sign in to view the complete answer paper.
Explain in brief about Multiplicative rules for weight tuning.
Answer
Sign in to view the complete answer paper.
What are the criteria for comparing learning algorithms?
Answer
Sign in to view the complete answer paper.
What are the approaches for deriving Confidence Interval?
Answer
Sign in to view the complete answer paper.
Define Instance based learning? What are the different types of instance based learning techniques?
Answer
Sign in to view the complete answer paper.
Explain in brief about Radial basis functions.
Answer
Sign in to view the complete answer paper.
Discuss in brief about Locally weighted Linear regression.
Answer
Sign in to view the complete answer paper.
What are the advantages and limitations of Locally weighted Linear regression?
Answer
Sign in to view the complete answer paper.
Discuss in brief about different search method for induction.
Answer
Sign in to view the complete answer paper.
Explain in brief about Genetic operators.
Answer
Sign in to view the complete answer paper.
Define Learning. Explain in brief about Models of evolution and learning.
Answer
Sign in to view the complete answer paper.
What is genetic programming? How can you represent Genetic programs?
Answer
Sign in to view the complete answer paper.