QBUS6850 Machine Learning for Business
商业机器学习代写 Question 1 (5 marks) From a bias-variance tradeoff perspective explain why bagged ensembles require stronger models than boosted ensembles.
(5 marks) From a bias-variance tradeoff perspective explain why bagged ensembles require stronger models than boosted ensembles.
(4 marks) Identify and describe two reasons why computer vision is challenging.
(5 marks) In the context of matrix factorisation, identify and outline a technique to estimate the factor matrices W and H.
Question 4 商业机器学习代写
(4 marks) In your own words, describe the cold start problem of recommendation systems and provide an example
(4 marks) In your own words, describe the purpose of bias units in a neural network
(6 marks) Name an example of a recommendation system that you have personally experienced and describe how you the recommendation system can be posed as a Multi-Armed Bandits problem.
(5 marks) Describe the operation of the Thompson Sampling Policy in the context of a Multi-Armed bandit model.
Question 8 商业机器学习代写
Suppose you are evaluating policies for the MAB environment with binary rewards.
Each bandit is Bernoulli distributed with the following parameters:
You have designed two policies and the action log is shown below:
Answer the following:
1. (6 marks) Select the policy which performs the best, explain your reasoning
2. (4 marks) Can you conclude that one policy is superior to the other based on this run?
(5 marks) Outline the steps of fitting an Adaboost model and match each step to the corresponding line/s in the code shown above.