Source for a short mathematical text by the author.
It introduces the theory of the multi-variate Gaussian (normal) probability distribution. It derives everything in an elegant, coordinate-free manner, using the full power of the modern theory of vector spaces (from linear algebra), unlike many existing textbooks on probability and statistics.
In less than 20 pages, the theory is developed, along with the essentials of the following widely-used applications (expressed in almost full generality):
- General linear regression
- Conditioning a Gaussian variable by another
- Kalman filtering
As far as the author knows, the article has no mistakes in the mathematical content, but it is still a draft that needs polishing.