Comments for Mixotricha

Comment on Extreme torture : do you chose the real thing or the memories ? by Arnaud

Arnaud — Wed, 11 Nov 2009 04:29:34 +0000

Hum… it’s almost going back to the traditional debate : “should we allow spanking [our kids] ?” with a certain twist. It may be a less radical example though i believe similar.

I’d like to quote “Mr. Dubois” on that : “… pain is the basic mechanism built into us by millions of years of evolution which safeguards us by warning when something threatens out survival. Why should society refuse to use such a highly perfected survival mechanism ?” – Starship Troopers, Robert A. Heinlein.

I think i’d go for A.

Comment on Making hidden patterns visible by reBlog from zyxo.wordpress.com: Mixotricha « Enyoing Digital Culture

reBlog from zyxo.wordpress.com: Mixotricha « Enyoing Digital Culture — Mon, 02 Nov 2009 21:14:38 +0000

[...] for you if you are not a “Homo analyticus”. Some examples :zyxo.wordpress.com, Mixotricha, Oct [...]

Comment on Link list for september 2009 by Link list for october 2009 « Mixotricha

Link list for october 2009 « Mixotricha — Sun, 01 Nov 2009 15:58:41 +0000

[...] this post ? Then you might be interested in the following : link list for september 2009 link list for august 2009 link list for juny 2009 link list for may 2009 Possibly related posts: [...]

Comment on The direction of evolution : speed matters ! by More Muslims are deciding that the discovery of Ardi disproves Darwin’s basic theory « My agnostic views & images I like

Sun, 01 Nov 2009 15:55:14 +0000

[...] The direction of evolution : speed matters ! (zyxo.wordpress.com) [...]

Comment on Data mining for marketing campaigns : interpretation of lift by Tim Manns

Tim Manns — Tue, 01 Sep 2009 02:50:25 +0000

Nice post.
I must admit I’m one of the lazy ones that often neglects to state the % of population when I quote lift.

In my case I usually go for 5% of population for a few superfical reasons;

– its 10 times higher than the incidence outcome I am predicting (lets say churn is 0.5%). I don’t want it too high or low, and magnitude of ten is a round number

– 5% of population gives us a reasonably good campaign size (and enough for valid control groups), especially after washing for privacy etc.

– our model predicts fairly well, so the model ‘drops off’ fairly quickly after 5% of population.

Cheers

Tim

Comment on evolution can occur in less than 10 years by Growth « Lary Crews: A Background Actor

Growth « Lary Crews: A Background Actor — Wed, 05 Aug 2009 17:10:55 +0000

[...] evolution can occur in less than 10 years (zyxo.wordpress.com) [...]

Comment on mathematics of information by zyxo

zyxo — Sat, 25 Jul 2009 19:44:41 +0000

John,
I get your point. What you mean is that both before and after receiving new information the information content is positive, as well as the information content of the newly received information. Meaning that adding somthing positive to something positive gets you something smaller than before.
Nevertheless : apparently quantum info can be negative

Comment on mathematics of information by John

John — Sat, 25 Jul 2009 19:38:24 +0000

Information content cannot be negative; the sum (or integral in the continuous case) of -p log p is always positive. However, in Bayesian statistics, it is possible for a particular observation to cause the posterior to be more diffuse than the prior, though on average the opposite is true. In a sense, this is a loss of information. Or maybe a loss of confidence in possible incorrect information. See this post for details.

Comment on Men Are More Accurate than Women… or lousy statistics ? by zyxo

zyxo — Fri, 17 Jul 2009 19:39:19 +0000

Will,
you are totally right. With statistics you never know for sure. The tests only say that the probability that the observed difference between the two groups (genders) are due to chance alone is less than 5%. Meaning that in 5% or 4% … of the cases you will see another outcome.
Compare it to what we know for sure : men are on the average taller than women. So take at random 4 men and three women and do the same exercise … I know a lot of women who are taller than a lot of men. And yet, statistics will tell you that men are taller than women.
About the other differences (education, age …) you could test these characteristics as well.
But allways remember that if you do 100 tests, you will find 5 or 4 or 3 or… of them that will be statistically significant, just by chance and hence give you false information.
Zyxo

Comment on Men Are More Accurate than Women… or lousy statistics ? by Will Dwinnell

Will Dwinnell — Fri, 17 Jul 2009 10:00:52 +0000

How would we know that it was their sex, specifically, which was the determining difference in performance between these two very small (!) groups?

For any 7 people chosen at random, surely there are many differences between a subset of 4 of them and the remaining 3, such as: their occupations, experience wielding hammers, education, age, race, health, visual acuity, general physical condition and so forth.