LogFAQs > #903493270

LurkerFAQs, Active DB, DB1, DB2, Database 3 ( 02.21.2018-07.23.2018 ), DB4, DB5, DB6, DB7, DB8, DB9, DB10, DB11, DB12, Clear

Topic List	Page List: 1
Topic	I've theorized that conservatives are on the losing side
COVxy 06/18/18 10:30:32 AM #34:	scar the 1 posted... COVxy posted... scar the 1 posted... Regression is a very important tool in statistical analysis - you should know this COVxy ~~:^)~~ http://phdcomics.com/comics.php?f=1921 I'm too lazy to try this, but what happens if you use the inverse distance as the norm instead? Shouldn't that produce something like the right panel? No idea, my intuition says no though. That you'd get a fit with the same slope but an infinitely large intercept. My guess is you'd need to do something clever with dot products to find the line that runs orthoganal to least squares fit. --- =E[(x-E[x])(y-E[y])] <cite>COVxy posted [p:903493270] in I've theor... [t:76718428]...</cite> <quote><cite>scar the 1 posted...</cite> <quote><cite>COVxy posted...</cite> <quote><cite>scar the 1 posted...</cite> <quote>Regression is a very important tool in statistical analysis - you should know this COVxy :^)</quote> http://phdcomics.com/comics.php?f=1921</quote> I'm too lazy to try this, but what happens if you use the inverse distance as the norm instead? Shouldn't that produce something like the right panel?</quote> No idea, my intuition says no though. That you'd get a fit with the same slope but an infinitely large intercept. My guess is you'd need to do something clever with dot products to find the line that runs orthoganal to least squares fit. --- =E[(x-E[x])(y-E[y])]</quote> ... Copied to Clipboard!
Topic List	Page List: 1