• Nie Znaleziono Wyników

Probability Calculus Anna Janicka

N/A
N/A
Protected

Academic year: 2021

Share "Probability Calculus Anna Janicka"

Copied!
14
0
0

Pełen tekst

(1)

Probability Calculus Anna Janicka

lecture X, 23.12.2020

CONDITIONAL EXPECTATION LINEAR REGRESSION

(2)

Plan for Today

1. Conditional expectation 2. Linear regression

(3)

Conditional Expectations

1. Intuition

2. Definition in the discrete case:

(4)

Conditional Expected Value of discrete RV

3. Example:

double 0-1

function of X

4. Transformations

(5)

Conditional density

5. Definition

(6)

Conditional density – cont.

6. Properties:

density

corresponds to conditional probability

different functions possible

OK for independent variables

7. Examples:

uniform distribution over square

“chain rule”

(7)

Conditional Expected value of continuous RV

8. Definition

9. Example

(8)

Conditional Expected value of continuous RV – cont.

10. Transformations:

(9)

Conditional Expectation

11. General definition of conditional expectation

12. Examples

(10)

Properties of Conditional Expectations

13. Properties of expected values

(11)

Properties of Conditional Expectations – cont.

13. Specific properties

(12)

Conditional Probability

14. Definition

(13)

Linear regression

1. Best (in terms of average square deviation) linear approximation of variable Y with variable X, i.e. aX+b:

minimizes solution:

(14)

Conditional Expectation as an approximation

1. Theorem:

Cytaty

Powiązane dokumenty

INDEPENDENCE OF EVENTS BERNOULLI PROCESS..

Well-behaved transformations of continuous

CUMULATIVE DISTRIBUTION FUNCTION – cont., EXPECTED VALUE – INTRO... The definition of

 Expected value for discrete random variables – cont..  Expected value for continuous random

In reality, we frequently do not know the distributions of random variables, and work with samples

Example: joint distribution is more than the aggregate of marginal distributions.. No simple definitions

Two-dimensional normal RV with mean and a covariance matrix Q. Two-dimensional

Best (in terms of average square deviation) linear approximation of variable Y with variable X,