Trees

Question 1

what is a main issue with logsitic regression

Answer 1

coeffcicent indicaates effect of variable not how decesion is made

Question 2

decesions we make in real life are:

Answer 2

in sequential order like regression trees

Question 3

main benefit of regrssion trees

Answer 3

is easy to understand

Question 4

in trees we know directrion the variable effects the probability but cant tell

Answer 4

impact on y variable

Question 5

we split the data using the IV into

Answer 5

yes or no decisions

Question 6

trees doesnt assumes model is

Answer 6

linear

Question 7

adding more splits will

Answer 7

increase accuracy

Question 8

3 splits in the tree

Answer 8

3 decision tree levels

Question 9

terminal node is an

Answer 9

output not condition ex- will tell you color red or grey

Question 10

root node is

Answer 10

condition very correlated with y variable - most important one at top

Question 11

we can have 100% accuracy if keep adding splits with no errors but issue is

Answer 11

too many variables and leads to overfitting

Question 12

what is fix to overfitting issue

Answer 12

set lower bond on number of points in each subset

Question 13

each split divides points into

Answer 13

buckets

Question 14

if we set minimum bucket size = lower bound thenn we

Answer 14

wont split if points in the split is less than minimum bucket size

Question 15

buckets only tell you

Answer 15

SIZE OF SPLit not the outcome of the bucket

Question 16

if bucket size too large then

Answer 16

model is too simple and will have poor accuracy

Question 17

logistic regression bucket size equivalent to

Answer 17

observations and predicts most frequent # or the baseline model

Question 18

in classfication model use

Answer 18

majority vote

Question 19

categorical variables/ classficiation are discrete variables what would using continous variables be

Answer 19

take average of all numbers, most frequent -> ex how many cars sold

Question 20

regression trees are easy to understand issue is when using single tree

Answer 20

we will observe errors significsantlly higher

Question 21

single trees are unstable meaning

Answer 21

small change in data= different trees/ interprattion

Question 22

how to fix issue of trees

Answer 22

forest

Question 23

random forest generate multiple trees instead of 1 then

Answer 23

takes average to come up with predicted values. -> if 5 trees then need 5 different data sets

Question 24

how does forest select trees

Answer 24

randomly select rows with replacement (same row can be selected )

Question 25

why do we select rows randomly with replacement

Answer 25

because trees highly sensitive to small changes so we change data slighlty to make regression tree and then by taking an average reduces the variance in the oredicted value

Question 26

minimum bucket in forests is called

Answer 26

Node size

Question 27

other parameter in trees

Answer 27

number of trees

Question 28

default number of trees in r

Answer 28

500

Question 29

`more trees is better because

Answer 29

unbiased

Question 30

why would more trees = issue

Answer 30

computationally dificult

Question 31

dont have to worry about what in forests

Answer 31

overfitting