Print a scatter plot of truth versus predicted % male

[aronwc]

Matplotlib is good for this.

[cyril94440]

[cyril94440]

Best is with alpha = 0.1 :

[aronwc]

I was thinking more something like this (points instead of bars):

http://www.cawcr.gov.au/projects/verification/scatterplot.gif

On Tue, Apr 15, 2014 at 6:47 AM, Cyril Trosset notifications@github.comwrote:

[image: Plot]https://cloud.githubusercontent.com/assets/2829562/2706571/96c779c2-c493-11e3-94f9-7d3efa372a2b.png

— Reply to this email directly or view it on GitHubhttps://github.com/tapilab/ctrosset/issues/16#issuecomment-40472349 .

[cyril94440]

With alpha = 0.1

[cyril94440]

With another representation :

it seems that the predicted values are pretty good, but I still get a mean square error of : -161.830505549

[aronwc]

Cyril,

Can you please generate a graph where the x-axis is the predicted %male and the y-axis is the true %male?

-Aron

On Wed, Apr 16, 2014 at 5:27 AM, Cyril Trosset notifications@github.comwrote:

With another representation : [image: capture decran 2014-04-16 a 12 26 22]https://cloud.githubusercontent.com/assets/2829562/2718627/9dd401c2-c551-11e3-801a-9424ca0c9a43.png

it seems that the predicted values are pretty good, but I still get a mean square error of : -161.830505549

— Reply to this email directly or view it on GitHubhttps://github.com/tapilab/ctrosset/issues/16#issuecomment-40584336 .

[cyril94440]

Alpha = 1

Alpha = 0.1

[cyril94440]

Alpha = 0.02