Visualizing Bayes' theorem • Hacker News Board

Re: Visualizing Bayes' theorem

#11

Post by dws » Sat, May 02, 2009, 3:41 PM UTC

vinutheraj wrote: post #4

I have always visualized probability and conditional probabilities in terms of Venn diagrams, it is easier to understand stuff this way. I think this is _the_ best way to teach probability to a beginner. What do you think ?!

Considering the universe of beginners (B), and the subset who are visual thinkers (V), then unless |V|/|B| >= 0.5, then no.

More data is needed.

Re: Visualizing Bayes' theorem

#12

Post by vinutheraj » Sat, May 02, 2009, 4:10 PM UTC

jodrellblank wrote: post #9

visitor4rmindia wrote: post #8

Earlier quoted context omitted.

I think, for a majority of people, it would be an easier way to study probability. Visualization makes things much more understandable than algebraic formula derivation.

Visualization makes things much more understandable Does it? Are you http://lesswrong.com/lw/dr/generalizing_from_one_example/ ?

Yea maybe I am bit biased about my opinion because I understood probability easier using visual examples, but people who I have interacted with mostly also understand things easier in the visual form. So I think it maybe _the_ best method to teach probability because it will reach out to a majority of people, ofcourse, there might be people who might understand it better if explained in a different way, but that maybe a minority, so all the different methods can be tried out in school.

One more thing I would like to add is that everyone has a biased opinion of things based on their perception of the world and that's why I think there should be debates. Generalizing from examples and experiences is a natural method developed during evolution.

If a kid touches a hot plate and experiences pain, I think it is perfectly valid that it should generalize that to all other hot plates. :)

Re: Visualizing Bayes' theorem

#13

Post by sketerpot » Sat, May 02, 2009, 7:02 PM UTC

I don't actually remember the formula for Bayes' theorem. I just visualize it as shown in this article, and write down equations from there. For me, anyway, that is actually a more reliable way of working.

Re: Visualizing Bayes' theorem

#14

Post by tpyo » Sat, May 02, 2009, 7:22 PM UTC

dws wrote: post #11

vinutheraj wrote: post #4

I have always visualized probability and conditional probabilities in terms of Venn diagrams, it is easier to understand stuff this way. I think this is _the_ best way to teach probability to a beginner. What do you think ?!

Considering the universe of beginners (B), and the subset who are visual thinkers (V), then unless |V|/|B| >= 0.5, then no. More data is needed.

According to http://en.wikipedia.org/wiki/Kinesthetic_learning 15% of people are kinaesthetic learners.

According to http://en.wikipedia.org/wiki/Auditory_learning 20% of people are auditory learners.

Since the only remaining group is visual learners, 65% of people must be visual learners.

Re: Visualizing Bayes' theorem

#15

Post by sdp » Sun, May 03, 2009, 3:41 AM UTC

jodrellblank wrote: post #7

vinutheraj wrote: post #4

I have always visualized probability and conditional probabilities in terms of Venn diagrams, it is easier to understand stuff this way. I think this is _the_ best way to teach probability to a beginner. What do you think ?!

I am reminded of something I read in Isaac Asimov's book on astronomy, where he talked about his great idea for how to visualize the size of distant objects and how lamentable it was that people don't use his obviously superior method. I stared at the pages unable to comprehend how he thought the size of a penny held a mile away was easier to visualise than the alternative. Likewise Eliezer's "Intuitive" explanation…

I've always been confused about Bayesian probability, until I read that article. I am a heavily visual learner, so the diagrams made a lot of sense to me...I finally get it!!