📊
Lecture 22 - Statistical
Inference
Review
The normal distribution function is a
probability density distribution function
characterized by two parameters of the
population: the mean of μ = 0.0and
the standard deviation σ = 1.0. Standard Normal Distribution
The use of the word “density” tells
us that it involves continuous
variables.
If xis a continuous variable, with a
normal probability distribution function,
the probability of a value of x < x0 us
the cumulative probability P (x < x0 )
The cumulative probability
distribution function CDF gives the
value of the cumulative probability
P (x < x0 )for any choice of x0 .
The cumulative probability for a
normal distribution is provided in the
table on D2L.
CDF
If a random variable xhas a normal
distribution with a mean μand a
Lecture 22 - Statistical Inference 1
standard deviation σ , the cumulative
distribution values can be found by
x −μ
calculating z0 = 0σ
Using the table: P (x < x0 ) =
P (z < z0 )
Other probabilities can be calculated
from this value by using:
P (x > x0 ) = 1 − P (x < x0 )
P (x0 < x < x1 ) = P (x <
x1 ) − P (x < x0 )
The central limit theorem CLT states that if xis a random variable and xis the
sample average, in a sample of nvalues, as nincreases, the PDDF for x
approaches a normal distrubution with a mean equal to the population mean of
xμx = μ
The standard deviation of the distrubution of sample means σx is the
standard deviation of the distribution for x, σ , divided by the square root
of the size n: σx =
σ
n
.
Mathematically fromt his equation, if you increase nin this equation, σx
decreases.
Confidence Limits
Supposed we wish to estimate a population mean μby taking a sample from
the population and calculated the sample average.
The sample average xis a “single point” estimator of the population mean.
Again, suppose we wish to estimate the population mean with a certain
degree of confidence (i.e. how confident should I be that my sample
avergae represents the population?).
If the sample size nis sufficiently large, the average and standard deviation of
all possible samples of a given size nhas an approximately normal distribution
Lecture 22 - Statistical Inference 2
with μx = μand a standard deviation for all possible samples of size nof
σx =
σ
n
.
Note that the probability of obtaining a value of xin the interval [μx −
2σ 2σ
2σx , μ + 2σx ] = [μ −
n
,μ+ n
]
Lecture 22 - Statistical Inference 3