Pearson Correlation

#interview #statistics #need-review

Extension of Co-Variance
Correlation tells us how strongly two random variables are related to each other.
Pearson Co-relation assume that both the datasets are from normal distribution, so it can't be used with discrete variables.

Pearson Co-relation is not affected by scale and so easy to interpret

Ranges from $[- 1, 1]$
We can interpret as,
- $- 1$ as totally Negative relationship
- $+ 1$ as totally Positive relationship
- $\approx 0$ as no relationship
Pearson-correlation depends on Variance and Co-Variance

Formula of Pearson Co-relation

$C o - r e l a t i o n (x, y) = \frac{C o - V a r i a n c e (x, y)}{\sqrt{v a r i a n c e (x)} \sqrt{v a r i a n c e (y)}}$

As we can see co-relation is dependent on Co-Variance and Variance

TODO:

Cosine Similarity vs Pearson Correlation

Related Notes