stats_Chapter 7


View on GitHub | Download Local

Click to view slide text

Chapter 7 Dr Wayne Stewart

Calculate n

Must know this proof for the exam

Method of moments to create point estimates

The following will be an example of MLE

Learn by doing!!

Go through all examples

Maximum likelihood method

Important!!

Maximum likelihood estimators

3 ways we will learn to make MLE’s

x

10

5.2

2

NR approximation

2.98

3 ways to perform MLE

0

20

40

f(x)

60

80

100

Newton-Raphson Algorithm

Newton Raphson derivation

Other methods

Other point estimators not discussed in detail within this course

JACKNIFE

ROBUST

JACK-KNIFE method

JackknifeR

R package

• Leave one(or more) datum out then estimate

Function that creates a vector of stats

compare

BOOTSTRAP

Interval estimates

Pivotal method

• Follow the proof

Two ways of obtaining ci’s

R as calculator

t.test()

n-1 degrees freedom

Two interpretations of ci’s

1

2

names(cc) head(cc) ccnewwithin(cc, ASPHALT factor(ASPHALT)) with(ccnew, boxplot(WATERLOSS ~ ASPHALT)) with(ccnew,var.test(WATERLOSS~ASPHALT)) with(ccnew, t.test(WATERLOSS ~ ASPHALT, mu = 0, var.equal = TRUE))

Welch 2 sample t-test

paired samples

Large sample ci for a proportion

ci for difference in proportions

ci for a population variance

The F pivotal statistic

Properties of the f statistic

𝐹𝛼 𝜈1 , 𝜈2 2

see proof

1

𝐹1−𝛼 𝜈2 , 𝜈1 2

ci for ratio of population variances

Calculate n for a half width

Check out all examples

Solution

We could use 𝑡𝛼 2

Not much difference The use of 𝑍𝛼 2

instead of 𝑡𝛼 is to 2

simplify the computation

PET EXAMPLE

R!!

Using the Bayesian paradigm

Bayes’ Rule

𝒑 𝜽 𝒇(𝒙|𝜽) 𝒑 𝜽𝒙 = 𝒑 𝒙

Use