Statistical Inference Lecturer: Prof. Duane S. Boning 1

advertisement
Statistical Inference
Lecturer: Prof. Duane S. Boning
1
Agenda
1. Review: Probability Distributions & Random Variables
2. Sampling: Key distributions arising in sampling
•
Chi-square, t, and F distributions
3. Estimation:
Reasoning about the population based on a sample
4. Some basic confidence intervals
•
•
•
Estimate of mean with variance known
Estimate of mean with variance not known
Estimate of variance
5. Hypothesis tests
2
Discrete Distribution: Bernoulli
• Bernoulli trial: an experiment with two outcomes
• Probability mass function (pmf):
f(x)
¾
¼
p
1-p
0
1
x
3
Discrete Distribution: Binomial
• Repeated random Bernoulli trials
• n is the number of trials
• p is the probability of “success” on any one trial
• x is the number of successes in n trials
4
Binomial Distribution
Binomial Distribution
1 .2
1
0 .8
0.25
0 .6
S eries1
0 .4
0 .2
29
27
25
23
21
19
17
15
9
13
7
11
5
3
1
0
0.15
0.1
0.05
29
27
25
23
21
19
17
15
13
11
9
7
5
3
0
1
Probability
0.2
Number of "Successes"
5
Discrete Distribution: Poisson
• Mean:
• Variance:
• Example applications:
– # misprints on page(s) of a book
– # transistors which fail on first day of operation
• Poisson is a good approximation to Binomial when n is
large and p is small (< 0.1)
6
Poisson Distributions
Poisson Distribution
0.2
0.18
=5
0.16
Probability
0.14
0.12
0.1
c
0.08
Poisson Distribution
0.06
0.1
0.04
0.09
=20
0.02
0.08
0
1
3
5
7
9
11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
0.07
Poisson Distribution
0.08
Probability
Events per unit
0.06
0.05
c
0.04
0.03
0.07
=30
0.06
0.02
0.01
Probability
0.05
0
1
0.04
3
5
7
9
11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Events per unit
0.03
0.02
0.01
0
1
3
5
7
9
11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Events per unit
7
Continuous Distributions
• Uniform Distribution
• Normal Distribution
– Unit (Standard) Normal Distribution
8
Continuous Distribution: Uniform
• probability density function (pdf)†
a
b
x
a
b
x
• cumulative distribution function* (cdf)
1
†also
sometimes called a probability distribution function
*also sometimes called a cumulative density function
9
Standard Questions You Should Be Able To Answer
(For a Known cdf or pdf)
• Probability x less than or
equal to some value
• Probability x sits within
some range
1
a
b
x
a
b
x
10
Continuous Distribution: Normal (Gaussian)
• pdf
1
• cdf
0.99865
0.977
0.84
0.5
0.00135
0.16
0.0227
0
11
Continuous Distribution: Unit Normal
• Normalization
• Mean
• Variance
• pdf
• cdf
12
Using the Unit Normal pdf and cdf
• We often want to talk about
“percentage points” of the
distribution – portion in the
tails
1
0.9
0.5
0.1
13
Philosophy
The field of statistics is about
reasoning in the face of
uncertainty, based on evidence
from observed data
• Beliefs:
– Distribution or model form
– Distribution/model parameters
• Evidence:
– Finite set of observations or data drawn from a population
• Models:
– Seek to explain data
14
Moments of the Population
vs. Sample Statistics
Population
Sample
• Mean
• Variance
• Standard
Deviation
• Covariance
• Correlation
Coefficient
15
Sampling and Estimation
• Sampling: act of making observations from populations
• Random sampling: when each observation is identically
and independently distributed (IID)
• Statistic: a function of sample data; a value that can be
computed from data (contains no unknowns)
– average, median, standard deviation
16
SticiGui: Statistics Tools for Internet and Classroom
Instruction with a Graphical User Interface
http://stat-www.berkeley.edu/~stark/SticiGui
Sampling Demo
17
Population vs. Sampling Distribution
Population
(probability density function)
n = 20
Sample Mean
(statistic)
n = 10
Sample Mean
(sampling distribution)
n=2
n=1
18
Sampling and Estimation, cont.
•
•
•
•
Sampling
Random sampling
Statistic
A statistic is a random variable, which itself has a
sampling distribution
– I.e., if we take multiple random samples, the value for the statistic
will be different for each set of samples, but will be governed by
the same sampling distribution
• If we know the appropriate sampling distribution, we can
reason about the population based on the observed
value of a statistic
– E.g. we calculate a sample mean from a random sample; in what
range do we think the actual (population) mean really sits?
19
Sampling and Estimation – An Example
•
Suppose we know that the thickness of
a part is normally distributed with std.
dev. of 10:
•
We sample n = 50 random parts and
compute the mean part thickness:
•
First question: What is distribution of
•
Second question: can we use
knowledge of
distribution to reason
about the actual (population) mean
given observed (sample) mean?
20
Estimation and Confidence Intervals
• Point Estimation:
– Find best values for parameters of a distribution
– Should be
• Unbiased: expected value of estimate should be true value
• Minimum variance: should be estimator with smallest variance
• Interval Estimation:
– Give bounds that contain actual value with a
given probability
– Must know sampling distribution!
Confidence Interval Demo
21
Confidence Intervals: Variance Known
• We know , e.g. from historical data
• Estimate mean in some interval to (1- )100% confidence
•
•
Remember the unit normal
percentage points
Apply to the sampling
distribution for the
sample mean
1
0.9
0.5
0.1
22
Example, Cont’d
•
Second question: can we use knowledge of
distribution to
reason about the actual (population) mean given observed
(sample) mean?
n = 50
95% confidence interval,
= 0.05
~95% of distribution
lies within +/- 2 of mean
23
Reasoning & Sampling Distributions
• Example shows that we need to know our sampling
distribution in order to reason about the sample and
population parameters
• Other important sampling distributions:
– Student-t
• Use instead of normal distribution when we don’t know actual
variation or
– Chi-square
• Use when we are asking about variances
– F
• Use when we are asking about ratios of variances
24
Sampling: The Chi-Square Distribution
• Typical use: find distribution
of variance when mean is
known
• Ex:
So if we calculate s2, we can use
knowledge of chi-square distribution
to put bounds on where we believe
the actual (population) variance sits
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0
5
10
15
20
25
30
25
Sampling: The Student-t Distribution
• Typical use: Find distribution of average when
• For k ! 1, tk ! N(0,1)
• Consider xi ~ N( , 2) . Then
is NOT known
• This is just the “normalized” distance from mean (normalized
to our estimate of the sample variance)
26
Back to our Example
• Suppose we do not know either the variance or the mean in
our parts population:
• We take our sample of size n = 50, and calculate
• Best estimate of population mean and variance (std.dev.)?
• If had to pick a range where
would be 95% of time?
Have to use the appropriate sampling distribution:
In this case – the t-distribution (rather than normal
distribution)
27
Confidence Intervals: Variance Unknown
• Case where we don’t know variance a priori
• Now we have to estimate not only the mean based on
our data, but also estimate the variance
• Our estimate of the mean to some interval with
(1- )100% confidence becomes
Note that the t distribution is slightly wider than the normal
distribution, so that our confidence interval on the true mean is
not as tight as when we know the variance.
28
Example, Cont’d
•
Third question: can we use knowledge of
distribution to
reason about the actual (population) mean given observed
(sample) mean – even though we weren’t told ?
n = 50
t distribution is
slightly wider than
gaussian distribution
95% confidence interval
29
Once More to Our Example
•
Fourth question: how about a confidence interval on our
estimate of the variance of the thickness of our parts, based on
our 50 observations?
30
Confidence Intervals: Estimate of Variance
• The appropriate sampling distribution is the Chi-square
• Because 2 is asymmetric, c.i. bounds not symmetric.
31
Example, Cont’d
•
Fourth question: for our example (where we observed
sT2 = 102.3) with n = 50 samples, what is the 95%
confidence interval for the population variance?
0.045
0.04
0.035
0.03
0.025
0.02
0.015
0.01
0.005
0
0
10 20
30 40
50 60 70
80 90 100
32
Sampling: The F Distribution
• Typical use: compare the spread of two populations
• Example:
– x ~ N( x,
– y ~ N( y,
– Then
from which we sample x1, x2, …, xn
2 ) from which we sample y , y , …, y
y
1
2
m
2
x)
or
33
Concept of the F Distribution
• Assume we have a normally
distributed population
• We generate two different
random samples from the
population
• In each case, we calculate a
sample variance si2
• What range will the ratio of
these two variances take?
) F distribution
• Purely by chance (due to
sampling) we get a range of
ratios even though drawing
from same population
Example:
• Assume x ~ N(0,1)
• Take samples of size n = 20
• Calculate s12 and s22 and take ratio
• 95% confidence interval on ratio
Large range in ratio!
34
Hypothesis Testing
• A statistical hypothesis is a statement about the parameters of a
probability distribution
• H0 is the “null hypothesis”
– E.g.
– Would indicate that the machine is working correctly
• H1 is the “alternative hypothesis”
– E.g.
– Indicates an undesirable change (mean shift) in the machine
operation (perhaps a worn tool)
• In general, we formulate our hypothesis, generate a random
sample, compute a statistic, and then seek to reject H0 or fail to
reject (accept) H0 based on probabilities associated with the
statistic and level of confidence we select
35
Which Population is Sample x From?
• Two error probabilities in decision:
– Type I error: “false alarm”
– Type II error: “miss”
– Power of test (“correct alarm”)
Consider H0 the
“normal” condition
• Control charts are hypothesis tests:
Consider H1 an
“alarm” condition
Set decision point (and
sample size) based on
acceptable , risks
– Is my process “in control” or has a significant change occurred?
36
Summary
1.
2.
Review: Probability Distributions & Random Variables
Sampling: Key distributions arising in sampling
•
3.
4.
Estimation: Reasoning about the population based on a
sample
Some basic confidence intervals
•
•
•
5.
Chi-square, t, and F distributions
Estimate of mean with variance known
Estimate of mean with variance not known
Estimate of variance
Hypothesis tests
Next Time:
1. Are effects (one or more variables) significant?
) ANOVA (Analysis of Variance)
2. How do we model the effect of some variable(s)?
) Regression modeling
37
MIT OpenCourseWare
http://ocw.mit.edu
2.854 / 2.853 Introduction to Manufacturing Systems
Fall 2010
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Download