Regression Model

Question 1

Training set

Answer 1

Data used to train the model

Question 2

x =

Answer 2

“input” variable feature

Question 3

y =

Answer 3

“output” or “target” variable

Question 4

m =

Answer 4

number og training examples

Question 5

m =

Answer 5

number of training examples

Question 6

f

Answer 6

the model

Question 7

prediction

Answer 7

estimated y

Question 8

linear function

Answer 8

f(x) = wx + b

Question 9

Univariate linear regression

Answer 9

One variable linear regression

Question 10

Cost function

Answer 10

Tells us how well the model is doing

Question 11

Cost function

Answer 11

Tells us how well the model is doing

Question 12

Goal of cost function

Answer 12

Minimize J(w,b)

Question 13

Gradient descent

Answer 13

Cost function
Start up and go down to try to get J to be smaller J(w,b)

Question 14

Important about gradient descent implimentation

Answer 14

Simultaneously update w and b

Question 15

If learning rate is to small

Answer 15

You take very small baby steps, so you do decrease w and b but,
Gradient descent may be slow.

Question 16

If learning rate is to large

Answer 16

You might miss your spot, so the cost might get worse

Gradient descent may : Overshoot, never reach mininum, fail to converge

Question 17

convex function

Answer 17

global minimum not local
bowl shape function

Question 18

Batch gradient descent

Answer 18

Each step of gradient decent uses all the training examples