Workshop — The Geometry of Statistics
I’m excited to present The Geometry of Statistics tonight, a new workshop for teachers. This workshop is about one of the coolest things I have learned over the past few years teaching linear algebra and writing a book on statistics: Finding the line of best fit for a set of data is really a geometry problem, but not the geometry problem you might think it is!
In this workshop we’ll see how finding the regression line is equivalent to finding the shortest path from a point to a plane in a curious high dimensional space. This geometric context helps make sense of many mysterious things in regression, like mean-centering (thanks, All 1s vector!) and the correlation coefficient. It also ties into advanced applied mathematical ideas like Principal Component Analysis (PCA) and the Singular Value Decomposition (SVD).
As is often the case when I learn mathematics, the turning point occurred when I finally understood why I didn’t understand. I look forward to sharing that understanding with others!
Related Posts
0 Comments