Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 5

TOPIC 3: PROBALISTIC MODELS

Probabilistic Model are commonly used to test the quality attributes, such as reliability,
availability, safety and performance of software-intensive systems by providing rightful
statistical result. Its accuracy depends on number of system properties or variables gathered
which will be estimated to make it clean and uniform. Environmental factors or system usage are
should be consider in estimating the its properties before doing any further processes.
That is why uncertainties takes place on in creating or doing such probabilistic models.
Scientists or researchers should account those uncertainties by solving them analytically or
through simulations. Inputted parameters are always assumed to be normally distributed as well
as reporting the mean and variances of the resulting attributes is usually considered sufficient.
However, in some cases some of uncertain factors doesn’t follow the normal distribution,
making analytical methods impractical to derive objective uncertainties and this will only make
our models complex in form.
Uncertainty is often called as error or bias in some cases, in world of statistics scientists
tries to quantify how much error that they might have in conducting their experiments, survey or
test results. Their goal is to make that error smaller as possible to make their final output reliable
and effective for further use. There are two main types of uncertainties, epistemic are things that
lacks of data or experience that makes it unknown for the researchers. Aleatoric on the other
hand are things that are simply unknown, like what number will show on the next roll.
Considering uncertainty in creating model allow scientist and researchers to provide a
precise interpretation based on gathered data. This will set a probable range that will tell the
acceptable percentage as well the percentage that should considered rejected. By incorporating
uncertainty explicitly in the model, we can measure the uncertainty associated with the outputs,
for example by giving a range to a forecast, which is more realistic goal.
In a business setting incorporating uncertainty is synonymous with understanding and
quantifying the risk in a business process, and ideally leads to better management decisions.

Measures of Uncertainty
Confidence Interval (CI) show what the uncertainty is with a certain statistic (like mean,
and etc.). The margin of error is a range of values above and below a confidence interval’s
sample statistic. For example, a survey might report a 95% confidence level of between 4.88 and
5.26. That means if the survey is repeated using the same methods, 95% of the time the true
population parameter will fall between 4.88 and 5.26, 98% of the time.

Sources of Uncertainty
 Interpolation errors happen because of a lack of data, and may be compounded by your choice
of interpolation method.
 Model bias happens because any model is an approximation, or a best guess at what a true
distribution might look like.
 Numerical errors are human errors that creep in when translating mathematical models into a
computer.
 Observational error is due to the variability of measurements in an experiment.
 Parameter uncertainty happens because we don’t know the exact, or “best” values in a
population—we can only take a good guess with sampling.
Error Propagation
Error propagation (or propagation of uncertainty) is what happens to measurement errors
when you use those uncertain measurements to calculate something else. For example, you might
use velocity to calculate kinetic energy, or you might use length to calculate area. When you use
uncertain measurements to calculate something else, they propagate (grow much more quickly
than the sum of the individual errors). To take this propagation into account, use one of the
following formulas in your experiments.

These formulas assume your errors are random and not correlated (e.g. if you have systematic
errors, you can’t use them).

1. Addition or Subtraction

Where:
 a,b,c are positive measurements
 x,y,z are negative measurements
 δ is the error associated with each measurement (the absolute error). δa is the uncertainty
associated with measurement a, δb is the uncertainty associated with measurement b, and
so on.
Example of Worked Formula
Let’s say you measured your height (a) as 2.00 ± 0.03 m. Your waistband (b) is 0.88 ± 0.04 m from the
top of your head, which means your pant length P would be p = H – w = 2.00 m – 0.88 m = 1.12 m.
The uncertainty, using the addition formula, is:

Giving a final measurement of 1.12 m ± 0.05 m.

2. Multiplication or Division formula


When calculating errors, there is no difference between multiplication and division.
3. Power formula
If n is an exact number and Q = xn, then

4. Measured Quantity Times Exact Number formula


If A is exact measurement (e.g. A = 9 or A = π) and Q = Ax, then:

δQ = |A| δx

5. General formula for Error Propagation


You might wonder why you can’t just add (or multiply, or divide) the errors and be done with it.
Why do we have to use formulas? Very basically, one small measurement error on an
independent variable, when applied to a function (say, a formula for area, kinetic energy, or
velocity) is going to result in a much larger error on the dependent variable.

These formulas are derived from the Gaussian equation for normally-distributed errors.
So, if you have some error in your measurement (x), then the resulting error in the function
output (y) is based on the slope of the line (i.e. the derivative).
The general formula (using derivatives) for error propagation (from which all of the other
formulas are derived) is:

Where Q = Q(x) is any function of x.

|DEFINITION of TERMS:
- Epistemic are things that lacks of data or experience that makes it unknown for the
researchers
- Aleatoric are things that are simply unknown, like what number will show on the next
roll.
- Confidence Interval (CI) show what the uncertainty is with a certain statistic (like mean,
and etc.)
- The Margin of Error is a range of values above and below a confidence interval’s
sample statistic.
- Error Propagation (or propagation of uncertainty) is what happens to measurement
errors when you use those uncertain measurements to calculate something else.

|ASSIGNMENT:
 In a paragraph, discuss and differentiate the 5 Error Propagations.

|EXERCISES & QUIZ:


1. In your own words, define what Probabilistic Model is.
2. Give and discuss the 5 Sources of Uncertainty.
3. What is Error Propagation?
4. Write down the two main parts of uncertainty.
5. In your own words, what is uncertain?

|KEY TO CORRECTION:
1. Probabilistic Model are commonly used to test the quality attributes, such as reliability,
availability, safety and performance of software-intensive systems by providing rightful
statistical result
2. //this is based on their perspective.
3. Error propagation (or propagation of uncertainty) is what happens to measurement errors
when you use those uncertain measurements to calculate something else.
4. Epistemic & Aleatoric
5. Uncertain: not exactly known or decided; not definite or fix; having some doubt about
something; not definitely known.

| SOURCES:
 https://www.statisticshowto.com/error-propogation/
 https://link.springer.com/article/10.1007/s10270-012-0277-5
 https://www.statisticshowto.com/uncertainty-in-statistics/
 https://www.statisticshowto.com/probabilistic/

| ADDITIONAL REFERENCES:
 https://youtu.be/ToPrATT-ISo
 https://youtu.be/LjGQsVv4KfE
 https://youtu.be/BRgbTvrSO2I

You might also like