Week 6 DSE

Question 1

What do we do before we create a machine learning model?
q

Answer 1

Visualise data

Question 2

How to use box plots to tell which variables are important?

Answer 2

Look at median

must be far apart

Question 3

What must we do whenever we have a categoric variable?

Answer 3

convert to numeric

Question 4

What are the ranges of result from the logistic model?

Answer 4

Probability Between 0 and 1

Question 5

More often than not we care about the ______ and ______ of the slope ,b1.

Answer 5

sign

relative magnitude

Question 6

When to use t value and why? For linear and multi

Answer 6

z-values: glm . Cause you CANT use least squares method . You use maximum likelihood for logistic which follows normal distribution

t-value: not glm, cause you use least squares method

Question 7

What does glm stand for?

Answer 7

generalised LINEAR model

Question 8

What must you do in r program to specify use of logistic model?

Answer 8

glm(default ~ balance, data = Default, family = binomial)

NEED TO SPECIFY FAMILY= BINOMIAL

Question 9

How do i represent scaling a certian independent variable in R?

Answer 9

use I (Represents operations)

= glm(default ~ balance + I(income/1000) + student, data = Default, family = binomial)

Question 10

What is sensitivity?

Answer 10

measures a classifier’s ability to identify positive status

p(tested POSITIVE | total that are actually positive)

how good we are at identifying positive cases out of all that are actually posittive

Question 11

What is specificity?q

Answer 11

measures a classifier’s ability to identify negative status

p(tested NEGATIVE| total that are actually negative)

how good we are at identifying negative patients correctly

True negative

Question 12

What are false positive?

Answer 12

fraction of cases that are ACTUALLY NEGATIVE, wrgonly classified as POSITIVE

Question 13

What is true positive?

Answer 13

Fraction of cases that are ACTUALLY POSITITVE, that are correctly classified as POSITITVE

sensitivity

Question 14

What happens as decision threshold increases?

Answer 14

FPR decreases. TPR decresaes

Question 15

What do we use to measure optimal decision threshold?

Answer 15

Draw multiple ROC curve at differnet thresholds

Find the one with the biggest area under the curve (AUC).

Question 16

What are on the axes on ROC curves?

Answer 16

Y axis: True positive

X axis: False posititve

Question 17

What is the range of values for AUC?

Answer 17

0 to 1

Best one is the one closest to 1

Question 18

What does model validation require?

Answer 18

Use another test dataset to see whehter it is good

split into train and test set

Question 19

How to measure the accuracy of model?

Answer 19

Training error

Question 20

What does test error quanitfy?

Answer 20

quantifies the predictability of the model.

Question 21

How does training error compare to test error?

Answer 21

training error - overly optimistic (UNDERESSTIMATE THE TEST ERROR)

because the model has already seen the data already,cannot predict new data

Question 22

What are the approaches to model validation?

Answer 22

validation-set approach

K-fold cross validation

Leave-one-out cross validation

Question 23

What is the most commonly applied model validation method?

Answer 23

k-fold cross validation

Question 24

GENERALLY, what are modle validation approaches like?

Answer 24

holding out subset(s) of the training observations from the model fitting process, and then applying the classifier to these held out observations.

Question 25

What is a disadvantage of validation-set approaches?

Answer 25

There is a randomization part here, wher ethe erorr rate will be differnent if you divide your data differently

Question 26

Explain what is validation-set approach?

Answer 26

(randomly divide the full data set into 2)

Question 27

Explain what is the k-fold cross validation approach

Answer 27

1. Randomly split the full data set into K folds of equal size. 2. training set: k-1 folds, test set: 1 fold 3. Iterate the process k times and then calculate average test erorr

Question 28

Explain leave one out cross validation

Answer 28

when k=n for a in range (n): test set= 1 obv training set= n-1 obv

Question 29

How to predict multiple obv in R?

Answer 29

need to use the c() function df_new = data.frame(student = c("Yes", "No"), balance = c(1500, 1500), income = c(40000, 40000)) predict(glm_fit, newdata = df_new, type = "response")

Question 30

What does type="response" mean?

Answer 30

return predicted probabilities

Question 31

How to generate confusion matrix in R?

Answer 31

table(glm_prob >= 0.5, Default$default) where glm_prob is the model AFTER PREDICTION

Question 32

how to write code for k fold cross validation k=5?

Answer 32

set.seed(1101) cv.glm(Default, glm_fit3, K = 5)$delta[1 Test error

Question 33

How to write leave one out cross valdiation?

Answer 33

cv.glm(Default, glm_fit3)$delta[1] # LOOCV

Question 34

How to get training error in R ?

Answer 34

glm_pred = ifelse(glm_prob_train > 0.5, "Yes", "No") mean(glm_pred != Default_train$default) # training error