Defense of Arguments: Bayesian Inference 1

Defense of this video:

http://www.youtube.com/watch?v=8Mylehb3NIU

Summary of video:

Predictions are more important than explanations, because you can have a reasonable sounding explanation for anything.

For example my explanation for randomess- is "the act of god interfering with nonrandomness"

Every creationist assumes that nature is random and God is nonrandom. I can just as easily declare the opposite, this is why explanations dont bolster the validity of the argument.

Explanations are important nly AFTER the validity has already been determined by predictions.




To creationists: Stop trying to explain evidence that already exists and start trying to predict some for the future.

Category:

Education

Tags:

• Evolution
• creationism
• intelligent design

[DarkMonkArt]

Filter out what you consider correct?

At lest half the time there correct ?

I will take the other half as being correct.

Sometimes they are not correct but you still fallow your Hypothesis as being correct.

The math you used shows exclusions of half the time being one way of the other?

SO that means Evolution 100 %?

Us just being here on earth with no (proven) contradiction 0%.

Hay What happen to the 50% of the time thing?

Begin your study with no opinion.

LIKE a Scientist should.

[truthofevolution]

what are you talking about? I dont understand what you are saying and you clearly dont understand what I said in the video.

[thevoicejam]

Thanks for your excellent argument. It seems the theory has bias without input from the opposing hypothesis. You stated that I.D has no testable hypothesis, wouldnt this then make it inappropriate for this stat inference? This would mean your argument losses its integrity. On the other hand if the opposing H can be tested you better input these as well.

[truthofevolution]

Hi what you say is technically correct that you cannot compare something with no testable hypothesis, however, while I say this for the layman, in utilizing the equation, we have to say that creationism accepts all possibilities with equal probability. This is the same as saying that it makes no hypothesis, but also allows for the utilization of this equation.

Thus there is input from the other side, which is all possibilities with equal probability. With evolution this is not the case.

[phicomingatya]

Also your coin experiment is a very bad and wrong example of how baysian theory works. A priori chances are multiplied by the likelihood of observed data. Even if only once, something else than head comes up, the chance of the theory 'head always comes up' will be multiplied by the likelyhood of zero, making he odds of the theory being true also zero. The theory will lose against any theory with positive likelihood.But please show me your magic equations turning zero chance into positive chance

[truthofevolution]

let me be a bit more clear: my whole point is the idea of comparison, and how something incomplete must be accepted if its better than the alternative: by saying "heads always comes up" Im implicitly setting up an alternative better than the one I have. Thats not my point.

I SHOULD have said that my theory says heads comes up without mentioning tails. Thus having the alternative of tails does not disprove it. This theory SUPERSEDES a theory which does not mention either heads or tails.

[phicomingatya]

Without mentioning tails, it's a totally different theory, that says head comes up (sometimes? not always?).

So this theory would be a hundred percent true no matter what the alternative. HUGE difference, still useless.

Again, with only one theory, baysian interference is useless.

Your science is useless as long as you do not make predictions for the theory that all of todays life is not all a result of evolviing.

You totally misrepresenting how baysian interference works.

[truthofevolution]

I already know about bayesian inference because I used it in my psych and genetics classes. Im referring to another video in this video. I compare evolution to creationism/intelligent design. Please watch my other video "proving evolution using predictions"

That video explains more clearly.

[truthofevolution]

"The theory will lose against any theory with positive likelihood."

Thats my whole point, there was no other theory to compare it too, thus we have to accept it. Until there is a to compete with it, we must accept it even though it is only partially correct.

I was not illustrating bayesian inference per se since that requires two theories, just that weaker theories can be superseded by a better theory.

Thats my whole point.

[phicomingatya]

Harsha, who taught you about baysian interference?

Don't you know that baysian theorem is useless with only one theory???

Second, with every observation/data, all theories/models are evaluated, depending on how likely the observation/data is given each theory.

Of course you already know all of this or you should.

Now I wonder if you can show me your equations and then I'll be happy to show you equations proofing evolution to be wrong, irrefuably!!!