Black Swans

simon

18 points

joined
Hey. I work for a sports betting company, and am starting to make some models of my own to predict tennis matches. πΎ
History

Recent History

Great idea - count me in!

Which GPU to Get for Deep Learning?

Hey,

I recently stumbled across a brilliant article explaining why you should invest in a GPU for training deep learning models and which one you should get.

I thought I'd share it here for those who are interested!

http://timdettmers.com/2018/08/21/which-gpu-for-deep-learning/

Hope you find it useful!

Which normality test is most powerful?

Hi,

As part of some research, I have to test whether a sample of observations is normally distributed. I've read about the Kolmogorov-Smirnov test, Jarque-Bera test and the chi-square test. Which of these (or another) is most powerful?

Thanks!

I'd be interested - sounds great!

Very interesting thoughts David. I'd also highly recommend both of these TED talks, especially the second by CΓ©dric Villani! Very inspiring.

Has Anybody Built a Neural Network to Predict Pool π± Matches?

Hey,

This is a bit random π€£, but I was wondering whether anybody on the site had ever used a neural network to predict pool or snooker matches? I'm trying to implement something myself, and am looking some advice.

Thanks

Here's my pythonic way of doing it:

1. Create a list of 5 foxes and 7 hounds (i.e., [FFFFFHHHHHHH]) and randomly shuffle it.
2. For each letter in the shuffled list, check whether it is an F followed by an H. If it is, add $1$ to a variable called successes.
3. Repeat this process $1000$ times.
4. Divide successes by $1000$ to get the average, which is approximately $2.91$ foxes are followed by a hound.

Thanks - very interesting!

Thank you so much. This is such a great explanation! I hadn't even heard of the central limit theorem before now - it's a pretty amazing result. It baffles me how you'd go about proving something like that.

Why does the normal distribution crop up everywhere?

Hi,

I haven't done much statistics up till now, but it seems like the normal distribution π is used all the time. Why is this? Is it for simplicity, or because it often provides a good fit for the problem?

Thanks.

I can't understand why it isn't just $n=2$ π£π£

To contact simon, email .