Image talk:H5n1 spread (with regression).png

From Wikipedia, the free encyclopedia

The colors in this image aren't chosen very well - I can guess that the regression is the smooth curve, but I can't visually differentiate them. Possibly due to my minor colorblindness? If someone who can easily alter this image sees this, it'd be nice to make the lines or colors a bit more differentiable. NegativeK 05:25, 11 January 2007 (UTC)

I'm not color blind and they look the same to me too. Was just coming here to point that out. Recury 17:24, 25 January 2007 (UTC)

Same Here, And I'm not color blind. Flubeca 22:54, 20 February 2007 (UTC)

[edit] Regression: not appropriate?

The regression plot brings a nice twist into the graph but it isn't particularly well pulled off and can't possibly be realistic. It seems that the intended objective was to model an exponential curve from the given data on flu deaths. That hypothesis doesn't make sense because it assumes that the deaths value can grow unrestrained, even outgrowing the number of infections. That, obviously, is not possible.

So in the end a more fitting regression method would be, for example, the logistic regression. As it accounts for a "ceiling", it can better fit that kind of values. So why not give it a try? --Mecanismo | Talk 01:00, 28 January 2007 (UTC)

I would leave out the regression entirely - the choice of which regression to use would probably be "unpublished synthesis", which is not allowed by WP:ATT. —AySz88\^-^ 05:20, 6 March 2007 (UTC)