- Distribution reflects a STRONG PATTERN in the DATA!!
- In some cases, Patterns are more important to study then DATA itself.
2: What is Distribution?
- Cumulative Distribution Function (CDF)
- Probability Mass Function (PMF)
- Probability Distribution Function (PDF)
--------------------------
Common Distributions and When to USE:
-------------------
3. Normal Distribution
When?
- If a variable depends on a large number of independent non-dominating parameters, the variable tends to follow Normal (Gaussian) Distribution. e.g
- Measure noise
- Human heights
- Exam grades
What is like? [Two parameter distribution]
Standard Normal Distribution: N(0, 1)
4. Uniform Distribution
When?
- Each value in X has the same possibility.
5. Exponential Distribution (Continuous)
When?
- model the waiting time for the next event to happen, e.g.
- the next vehicle passing a line
- bank service time duration
What?
6. Geometric Distribution (Discrete)
When
- Fail times util success Once, similar with Exponential
What
7. Binomial distribution (Discrete)
When
- Try N times and success K times
- When N -> INFINITY, it becomes Normal
What
8. Poisson Distribution
When
- Counts of random events, e.g.
- number of phone calls received one day
What
9. Chi-2 Distribution
Definition
X1^2 + X2^2 + X3^2 + ... + Xk^2
Given,
Xi ~ N(0, 1)
When
Testing
What
10. F-Distribution
When
What
11. T-Distribution
No comments:
Post a Comment