0% found this document useful (0 votes)
37 views13 pages

ECE3530 Slides Introduction

The document discusses applications of probability and statistics including cell phone transmission over noisy mediums, weather prediction, speech recognition, and image compression. It then uses an example of testing a new drug for Alzheimer's disease, which involves measuring ventricular volume from MRI scans of patients in a clinical trial receiving either the drug or a placebo. The data from the trial is analyzed visually with plots and quantitatively by calculating sample means and variances to determine if the drug was effective in slowing the progression of Alzheimer's disease. Descriptive statistics provide a summary of the data, while statistical inference is used to make higher-level decisions about the data.

Uploaded by

Natnael Temesgen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
37 views13 pages

ECE3530 Slides Introduction

The document discusses applications of probability and statistics including cell phone transmission over noisy mediums, weather prediction, speech recognition, and image compression. It then uses an example of testing a new drug for Alzheimer's disease, which involves measuring ventricular volume from MRI scans of patients in a clinical trial receiving either the drug or a placebo. The data from the trial is analyzed visually with plots and quantitatively by calculating sample means and variances to determine if the drug was effective in slowing the progression of Alzheimer's disease. Descriptive statistics provide a summary of the data, while statistical inference is used to make higher-level decisions about the data.

Uploaded by

Natnael Temesgen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

ECE

ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Applications of probability & statistics


• Cell phone transmission over a noisy medium
has a probability of error. Number of extra bits
sent depend on statistics of the noise.
• Weather prediction based on past
observations makes use of probabilities.
• Speech recognition is based on determining
the most likely (highest probability) spoken
words based on statistics of past speech. The
statistics may be speaker specific.
• Image compression standards like jpeg make
use of unequal probabilities of pixel intensities.
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Example: Testing the effect of a new drug


• Problem: A neurologist wants to determine if a
certain drug slows down the progress of
Alzheimer’s disease.
• It is known that Alzheimer’s disease results in
abnormal enlargement of the ventricles (a
compartment of the brain) as it progresses.

Alzheimer’s patient Normal brain


ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Measuring ventricular volume


• The neurologist works together with an
engineer who specializes in digital image
processing to develop a computer program that
automatically measures ventricular volume
from Magnetic Resonance Images (MRI)

MRI Partitioned image


ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

The clinical trial


• The neurologist then recruits 20 Alzheimer’s
patients into a clinical trial.
• He randomly assigns the 20 patients into 2
groups of 10. One group will take the drug that
is being tested for 6 months while the other
group will take the placebo (no drug) for 6
months.
• At the beginning of the 6 months all patients
have a MRI taken and their ventricular volume
is measured.
• This is repeated at the end of the 6 months
which allows us to compute the change in
ventricular volume for each patient.
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

The data
• The following ventricular volume changes (in
units of ml) are recorded
Drug 4.5 3.5 7.8 -1.1 5.8 7.2 6.7 6.2 4.6 6.5

Placebo 10.5 9.6 7.4 7.6 10.6 6.4 11.6 11.0 6.8 8.5

• Do you think the drug was effective in slowing


down Alzheimer’s disease?
• How can we start to look at this data…
– in a more visual way?
– in a more quantitative way?
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Looking at the data visually


• Lets make a plot of the data:

– Crosses correspond to “drug” sample


– Circles correspond to “placebo” sample
• Does this help in forming an opinion?
• Does it directly allow us to make a formal
decision on whether the drug is effective?
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Towards more quantitative analysis


• If you had to summarize both samples with a
single number what would you choose?
– Answer: Sample mean
• Lets take the “drug” sample:
Drug 4.5 3.5 7.8 -1.1 5.8 7.2 6.7 6.2 4.6 6.5

– What is the sample mean?5.17


• How about the “placebo” sample:
Placebo 10.5 9.6 7.4 7.6 10.6 6.4 11.6 11.0 6.8 8.5

– What is the sample mean? 9.0


ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Lets change our data a bit…


• Assume the following data was recorded
instead:
Drug 8.6 5.6 -0.4 6.7 5.9 11.9 10.2 1.0 -1.9 4.1
Placebo 7.3 6.7 12.3 18.2 6.0 4.2 7.8 13.3 9.1 5.1

• Do you think the drug was effective in slowing


down Alzheimer’s disease in this case?
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Compute the sample means


• “Drug” sample:
Drug 8.6 5.6 -0.4 6.7 5.9 11.9 10.2 1.0 -1.9 4.1
– Sample mean = 5.17
• “Placebo” sample
Placebo 7.3 6.7 12.3 18.2 6.0 4.2 7.8 13.3 9.1 5.1
– Sample mean = 9.0
• Observation: These are the exact same
sample means as before!
• Conclusion: Sample mean by itself is not
sufficient to describe the data.
• Question: So what changed?
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Measuring variability
• Sample variance is a measure of variability.
• Sample variances in our examples:
Units are in ml2

Drug Placebo There is greater


variability in the
Dataset 1 6.61 3.59 second dataset.
Dataset 2 20.55 18.99 Large variability can
hide the difference
• Sample standard deviations: between two samples.
So even though the
Units are in ml

Drug Placebo sample means


Dataset 1 2.57 1.89 remained the same,
we are less sure about
Dataset 2 4.53 4.36 what to conclude about
dataset 2.
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Descriptive and inferential statistics


• Mean, variance and standard deviation are
what we call descriptive statistics. There are
many more such as median, range, etc.
– Descriptive statistics provide a summary of our data.
• When we make higher-level decisions about
our data this is called statistics inference.
– For instance: We decide that the drug under testing is
not effective in slowing down Alzheimer’s disease.
– Descriptive statistics become important tools in
making statistical inferences.
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Probability and Statistics


• For a statistical problem, the sample along
with inferential statistics allows us to draw
conclusions about the population using
elements of probability.
• Problems in probability allow us to draw
conclusions about characteristics of
hypothetical data taken from the population
based on known features of the population.
ECE
ECE 3530
3530 notes
notes -- T.
T. Tasdizen
Tasdizen

Some other things…


• Observational study vs. experimental design
• Discrete vs. continuous data
• Random sampling, sample size
• Biased vs. unbiased sample

You might also like