Neural Net Lottery Picker

CUDA wrapper would be great, I am already set up and ready to go. By the way, I am noticing that the interface since version 2 is getting a little sluggish. Just thought I would mention it.

adamcustom,

There is a space at the end of each line causing the error message.

So first you have to remove all the spaces to fix this issue.

Good luck!

Sorry, Also forgot to mention that the program is now producing duplicate picks as well. I don't know if this is because of the new component. I went back to using v1.22.

goodlukman,

You can get duplicate picks when the 'Learning Rate' (alpha) value is to high.

I use a small value here like 9E-05 in combination with 150 to 300 neurons (and latest 10 to 60 draw results); these settings give me the best results (pick 6/45 game)

I'll give v2.0  another try today. Thx

Do you have short definitions for all the used terms, like noise for example?

Thanks for this new version!!! Stoopendal...

It take time to find the sucess parameters but its a fun thing to do it.

Thanks again.

Hi SergeM,

Noise: 

With this factor you manipulate the weight of the initial randomly generated output values.  I'm not sure what the impact of this factor is on the 'performance' of the network (haven't done enough testing).

Moment factor  meaning: 

1. This can improve the learning rate in some situations, by helping to smooth out unusual conditions in the training set.
2. Momentum simply adds a fraction m of the previous weight update to the current one. The momentum parameter is used to prevent the system from converging to a local minimum or saddle point.A high momentum parameter can also help to increase the speed of convergence of the system. However,setting the momentum parameter too high can create a risk of overshooting the minimum,which can cause the system to become unstable.A momentum coefficient that is too low cannot reliably avoid local minima, and can also slow down the training of the system.

Learing Rate meaning (source http://www.cheshireeng.com/Neuralyst/nnbg.htm):

The learning rate, LR, applies a greater or lesser portion of the respective adjustment to the old weight.
If the factor is set to a large value, then the neural network may learn more quickly, but if there is a large variability in the input set then the network may
not learn very well or at all. ...
Usually, it is better to set the factor to a small value and edge it upward if the learning rate seems slow.

The neural network I use in NNLP is a brute force approach of finding a lower average error for the  training data (by randomly generating the weights, etc. for each training).  However  I'm planning to implement epochs (training cycles) to the tool to get a lower error for each training vector (= previous draw result ==> next draw result in the training set) resulting in a much lower average error for the training set.

An epoch is a measure of the number of times the training vector is used once to update the weights

I'm really anticipating the epoch addition. This will put your program model in line with similar programs. Keep up the great work. Got another 3/5 today! This is by far the most consistent winning I've experienced the entire year. Thank you.

A couple of suggestions for features:

1. a pling/signal can be heard when a pick has been found.
2. decide how many picks you want and set the number, so it stops training when the number of picks has been reached. That way we can leave the program to work alone while we do other stuff.

Im trying something different with NM p3 evening im inserting 999 after each entry for tonight.

inputs

8 9 1
9 9 9
4 8 7
9 9 9
1 1 1
9 9 9
1 5 5
9 9 9
1 3 5
9 9 9
2 8 4
9 9 9
4 3 6
9 9 9
3 7 4
9 9 9

outputs:

1-02-03-;45.1013142180644;1.67893780070288;2.2619788514038;17:56:39.8183879;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50
08-03-06-;13.0938610416533;0.904636012495265;1.01645291895239;17:56:40.3652629;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50
07-03-06-;12.7987774426448;0.894384475583796;0.98920892172863;17:56:49.3652629;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50
06-05-04-;12.6965655235923;0.89080600874967;0.945950218515321;17:56:57.3183879;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50
05-07-06-;12.3878753627381;0.879910342120792;0.880808572025811;17:57:06.2402629;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50
06-07-05-;12.1174002478979;0.870251409360319;0.826951592535781;17:57:26.1465129;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50
05-07-08-;12.0834879127799;0.869032792562366;0.744719203111743;17:58:05.9590129;Alpha:9E-05| Noise:0.25|Threshold Error:0.01|Hidden Neurons:50|Test Data Range:50

thanks!

I am still trying to win the keno in Atlantic Canada!!

I will now try to predict picks based on draw order of the numbers on the lottery.

I don't have enough data to do a backtest, so I will create ten picks for the next saturday lotto draw.

I will use the best settings based on earlier backtesting on the saturday lotto.

I´m agree.Numbers and not only combinations predicted,would be a great feature!.

Stoopendal Thanks for your big efforts.

Regards