Hi AllStaters
I'm sending out the following email on behalf of a friend who has a new
found interest in Stats through gambling! My expertise on Neural Networks
and Data Mining is very limited.
His email is as follows:
I'm currently working on a spreadsheet that generates odds for 3 events
based on 4 different selection methods, say:
Event: A B C
Method 1 gives: 1.61 2.42 8.86
Method 2 gives: 1.42 3.05 9.50
Method 3 gives: 1.38 2.78 9.98
Method 4 gives: 1.50 2.86 9.23
I've made these numbers up, but in reality, taking reciprocals of the
numbers in each line and adding them together should always sum to 1.
I then get another 'true set' of odds, say:
True values: 1.53 2.92 9.10
I can get this data for several hundred cases, and I want to know what
weighting I should assign to each method, to be able to predict the true
odds. ie., which method is most accurate. I guess it's a neural network
problem to be honest, but I don't have the practical experience in dealing
with them to make it work.
Can anyone help or recommend books on the subject? Also please let me know
if you mind if I send him your email address with any responses, he may
want to contact you in future!
Many thanks
Eloisa Ruiz
GMS Graduate Statistician
GlaxoSmithKline, Cork
[log in to unmask]
Phone: +353 21 4512130
|