predictive analytics; prediction

Question 1

confidence interval

Answer 1

quantifies uncertainty surrounding mean over a large quantity

Question 2

prediction interval

Answer 2

quantifies uncertainty surrounding prediction of single quantity

Question 3

machine learning definition

Answer 3

study of algorithms applied to data, focussing on prediction and classification

Question 4

machine learning characteristics

Answer 4

• formulae like regression, but rules to produce yhat
• translated into computer code and automated
• predictions generated quickly and repeated without intervention
• takes error out of humans choosing variables
• challenges of interpretability and bias

Question 5

regression process

Answer 5

• hypothesis
• select variables
• train
• test
• select best model
• refine

Question 6

problems of machine learning

Answer 6

1) good for prediction, not explanatory
2) black box (don’t know why machine chooses variables)
3) ethical (sub-optimal outcomes like discrimination)

Question 7

black swan

Answer 7

extreme outliers that have a disproportional effect (overfitting)

Question 8

overfitting

Answer 8

• only see particular subset of data, and what is true is smth bigger we don’t observe
• trying so hard to explain what we see, so poor at explaining what we don’t see

Question 9

avoiding overfitting

Answer 9

• split data into training and testing
• training: subset of data, eg 80% to estimate formula
• testing: 20% of data to test how well model predicts variables not shown

Question 10

R squared

Answer 10

increases with number of variables

Question 11

adjusted R squared

Answer 11

• penalises for number of variables
• asks if I gain much explanatory power for extra variables
• can go down if irrelevant variable is included
• bigger is better = more accurate

Question 12

mean squared error

Answer 12

• common mean of selecting model
• smaller is better = less error
• best model = one with least errors
• high 𝑅^2 implies low MSE

Question 13

regression tree

Answer 13

• output from machine learning
• segments inputs in mutually exclusive/exhaustive regions
• branches connect nodes to a terminal node (leaf)