1/20
Flashcards about Covariance and Correlation
Name | Mastery | Learn | Test | Matching | Spaced |
---|
No study sessions yet.
Covariance
A measure of how two variables change together.
Covariance Measurement
Compares data points to a defined point (x, y) to measure rise and fall.
yi – y
Distance from y
xi - x
Distance from x:
(xi – x)(yi – y)
Individual distance from xy:
∑ (xi – x)(yi – y)
Total distance from xy:
∑ (xi – x)(yi – y) df
Covariance (x,y)
∑ (yi – y)(yi – y) df
Variance of y
Big math difference
Covariance can be negative
Z transformation to Scale Covariance Zxi = xi - x sx
Centers these data Zx = 0
Zxi is unitless
Scales these data Szx = 0
Correlation
The preferred measure of the rise and fall of (x, y) because it gets around the unit issue -1 ≤ ρ ≤ 1
ρ
Correlation is standardized Covariation
Correlation Question
Asks the question: “Is there a true correlation between x and y?”
H0
ρx,y = 0
HA
ρx,y ≠ 0
𝑥
parameter μ
r
parameter ρ
Pearson correlation coefficient
r is known as the Pearson correlation coefficient
Hypothesis testing for correlation
How can we determine the probability that our sample correlation (r) is drawn from the true null correlation (ρ= 0)?
t score
t score = r − ρ 𝑆𝑒𝑟