Advanced Machine Learning (ELEDSE7A) — DU B. Sc. (Hons) Electronics Previous Year Question Papers

Jnique Paper Code 2513010017• Yame of the Paper Advanced Machine Learning Instruction for Candidates Q1 (a) Define supervised learning. Give one example. 3 (b) What is the purpose of a cost function in regression? 3 (c) Explain the meaning of a cluster centroid in K-Means. (d) What is a Conditional Probability Table (CPT) in Bayesian 3 (e) State the meaning of a utility function in MDPs. (f) What is the mistake-bound model in learning theory? 3 (a) Discuss simple linear regression and explain how the line of best fit is 6 (b) Explain the working of a Decision Tree classifier. Discuss Gini 6. Describe the importance of feature scaling. Mention two scaling 6 Q3 (a) Explain the working of Agglomerative Clustering with an example. 6 (b) Describe the steps for computing principal components in PCA. 6 (c) Compare association rule mining with clustering. Give one 6 application for each. Q4 (a) What is a Bayesian Network? Explain how inference is performed. 6 (b) Differentiate between a Simple Markov Model and a Hidden Markov 6 (c) Explain value iteration and policy iteration in an MDP. Q5 (a) Explain the PAC learning framework. What does "probably 6 approximately correct" mean? (b) Describe the working principle of the Distance Weighted Nearest 6 Neighbour algorithm. (c) Discuss factors influencing sample complexity in computational learning theory. Q6 (a) Compare regression and classification with examples. 6 (b) Explain how DBSCAN identifies noise points in a dataset. 6 (c) Describe polynomial regression and mention when it is preferred. Discuss passive reinforcement learning with an example. Compare Random Forest and SVM with respect to leaming strategy and generalization.