Welcome Guest
Log In | Register )
You last visited December 6, 2016, 2:46 pm
All times shown are
Eastern Time (GMT-5:00)

Artificial Intelligence

Topic closed. 73 replies. Last post 6 years ago by martor854.

Page 1 of 5
51
PrintE-mailLink
themagician's avatar - artificial intelligence-1.jpg

France
Member #92435
June 7, 2010
43 Posts
Offline
Posted: June 14, 2010, 11:43 am - IP Logged

Hi,

I will like to know if somebody is seriously studying or using artificial intelligence for lotto, I know there are some old posts about it and some explanations but has somebody got good results using artificial intelligence software or algorithms?

Is important to me because I am currently working on this kind of software, using backpropagation and it will be very helpful to get some feedback from your experience.

Thank you ;-)

Imagination can take you anywhere

http://www.artificialintelligencelotto.com 

    RL-RANDOMLOGIC's avatar - usafce

    United States
    Member #59354
    March 13, 2008
    3967 Posts
    Offline
    Posted: June 16, 2010, 3:24 pm - IP Logged

    Hi,

    I will like to know if somebody is seriously studying or using artificial intelligence for lotto, I know there are some old posts about it and some explanations but has somebody got good results using artificial intelligence software or algorithms?

    Is important to me because I am currently working on this kind of software, using backpropagation and it will be very helpful to get some feedback from your experience.

    Thank you ;-)

    themagician

    If you find something then please pass it on. I too have been working with AI and NN but as yet have

    nothing to share. Have you ever looked into Brain Maker, if not then google it and read some of the

    examples.  I was most interested in the example of the binary srting prediction that was above 90%

    after being trained. It is a bit high $$$ but I might give it a shot.  backpropagation I think would be the

    way to go concerning the lottery but I think before any real headway could be made you would need

    to use other data instead of the numbers.  Think of it this way, the numbers are the final result of the

    lottery drawing. Try to reverse engineer the process looking for markers along the way. Much data

    would be needed to find high quality markets but I think it might lead to something.  I did find a old

    program written in basic years ago of a AI or NN that gave you two choices and asked you to press

    a key after you had decided. It would then make a prediction as to which choice you had made. Next

    you would press the key for correct or incorrect .  The more you played the better it got.  If I can find

    it again I will pass it on.  At the time I never had the time to check the code and see how it worked.

     

    RL

    Working on my Ph.D.  "University of hard Knocks"

    I will consider the opinion that my winnings are a product of chance if you are willing to consider

    they are not.  Many great discoveries come while searching for something else

    USAF https://en.wikipedia.org/wiki/Prime_Base_Engineer_Emergency_Force

      US Flag Trump / 2016 & 2020  

      Avatar

      United Kingdom
      Member #31295
      January 27, 2006
      73 Posts
      Offline
      Posted: June 17, 2010, 4:40 pm - IP Logged

      Found this link to AI programs and neural networks.

      http://www.dmoz.org/Computers/Artificial_Intelligence/Neural_Networks/Software/

       

      I have been 'tinkering' around with converting drawn numbers into binary patterns eg if the number 3 was drawn once over 5 draws the pattern may look like this F,T,F,F,F where T = drawn and F = not drawn. It is possible to find patterns that repeat and some are more frequent than others. But as always the problem knowing when the patttern will occur again.

        themagician's avatar - artificial intelligence-1.jpg

        France
        Member #92435
        June 7, 2010
        43 Posts
        Offline
        Posted: June 23, 2010, 9:21 am - IP Logged

        Ok, I will take a look at Brain Maker, a little bit expensive though..

        It is true you cannot input "raw" numbers to the network,and there are some aproaches here, either taking the whole thing as a 1-0 matrix, or to scale all values before input, ah, I keep on testing.. and I do everything on my sparetime, last time I've got 18-20-36.. the comming out numbers beeing 19-21-36.., and it is good, getting closer, I want to get some more

        Imagination can take you anywhere

        http://www.artificialintelligencelotto.com 

          RL-RANDOMLOGIC's avatar - usafce

          United States
          Member #59354
          March 13, 2008
          3967 Posts
          Offline
          Posted: June 24, 2010, 3:15 am - IP Logged

          Ok, I will take a look at Brain Maker, a little bit expensive though..

          It is true you cannot input "raw" numbers to the network,and there are some aproaches here, either taking the whole thing as a 1-0 matrix, or to scale all values before input, ah, I keep on testing.. and I do everything on my sparetime, last time I've got 18-20-36.. the comming out numbers beeing 19-21-36.., and it is good, getting closer, I want to get some more

          TM

          Here is what I done. I made 10 groups of 4 numbers for my 5-39.  Group 10 only had 3 numbers

          because it is a 5 of 39.  Next each time any of the numbers from a group hit it would count as a

          "1" if the group missed it would be a "0". each draw was converted into a binary string that looked

          like this.

          1 0 0 1 0 0 1 1 0 1   = 03-15-26-31-39

          1 1 0 1 0 0 1 0 0 0

          0 0 1 1 0 0 0 1 0 1

          1 0 0 1 0 0 0 0 1 1

          0 0 1 0 1 1 0 1 1 0

           

          I done this for all draws in my database.  Next I read the data from top to bottom placing

          each columb in a string.  I now had 10 strings of 0's and 1's the lenth of my drawfile.  I then

          searched each of the 10 long strings for patterns.  Many repeating patterns can be found

          I would like to run these strings using Brain maker.  Lets say that it said that the next bit

          in the first columb should be a "1" then I would know that the first number, if correct would

          be one of "01-02-03-04"  some draws would have more then 1 number that hit within a group

          so to make this more accurate one could shuffel the numbers in the groups and then rerun.

          I done much with this but only got about 60 to 70 percent hit rate for a couple of strings and

          much lower for the rest.  I used a program that I wrote to do this and did not use brainmaker.

          BrainMaker web page shows a sample for predicting the next digit of a random generated string 

          with a very high hit rate.  You could use 2 numbers in each group and have a 20 bit long string.

          or 3 numbers for a 13 bit string.

          RL

          Working on my Ph.D.  "University of hard Knocks"

          I will consider the opinion that my winnings are a product of chance if you are willing to consider

          they are not.  Many great discoveries come while searching for something else

          USAF https://en.wikipedia.org/wiki/Prime_Base_Engineer_Emergency_Force

            US Flag Trump / 2016 & 2020  

            themagician's avatar - artificial intelligence-1.jpg

            France
            Member #92435
            June 7, 2010
            43 Posts
            Offline
            Posted: June 25, 2010, 9:38 am - IP Logged

            TM

            Here is what I done. I made 10 groups of 4 numbers for my 5-39.  Group 10 only had 3 numbers

            because it is a 5 of 39.  Next each time any of the numbers from a group hit it would count as a

            "1" if the group missed it would be a "0". each draw was converted into a binary string that looked

            like this.

            1 0 0 1 0 0 1 1 0 1   = 03-15-26-31-39

            1 1 0 1 0 0 1 0 0 0

            0 0 1 1 0 0 0 1 0 1

            1 0 0 1 0 0 0 0 1 1

            0 0 1 0 1 1 0 1 1 0

             

            I done this for all draws in my database.  Next I read the data from top to bottom placing

            each columb in a string.  I now had 10 strings of 0's and 1's the lenth of my drawfile.  I then

            searched each of the 10 long strings for patterns.  Many repeating patterns can be found

            I would like to run these strings using Brain maker.  Lets say that it said that the next bit

            in the first columb should be a "1" then I would know that the first number, if correct would

            be one of "01-02-03-04"  some draws would have more then 1 number that hit within a group

            so to make this more accurate one could shuffel the numbers in the groups and then rerun.

            I done much with this but only got about 60 to 70 percent hit rate for a couple of strings and

            much lower for the rest.  I used a program that I wrote to do this and did not use brainmaker.

            BrainMaker web page shows a sample for predicting the next digit of a random generated string 

            with a very high hit rate.  You could use 2 numbers in each group and have a 20 bit long string.

            or 3 numbers for a 13 bit string.

            RL

            If I understand well you test if a column formed by a group of four digits is true or false for the next drawing. If your second group is true, then one or more of these numbers 5,6,7,8 should be comming...mmm..
            One of the approaches I had was taking the entire game as a matrix, or a string, in your game it will be a 1 to 39, lets's say then that number 3 has come out then the third position of my array will be 1 "true", and 0 "false" for all the other numbers that didn't came out. Well I've tested a lot with this approach and yet no good conclusion..

            Imagination can take you anywhere

            http://www.artificialintelligencelotto.com 

              RL-RANDOMLOGIC's avatar - usafce

              United States
              Member #59354
              March 13, 2008
              3967 Posts
              Offline
              Posted: June 27, 2010, 12:37 am - IP Logged

              If I understand well you test if a column formed by a group of four digits is true or false for the next drawing. If your second group is true, then one or more of these numbers 5,6,7,8 should be comming...mmm..
              One of the approaches I had was taking the entire game as a matrix, or a string, in your game it will be a 1 to 39, lets's say then that number 3 has come out then the third position of my array will be 1 "true", and 0 "false" for all the other numbers that didn't came out. Well I've tested a lot with this approach and yet no good conclusion..

              TM

              Thats the idea. Hope you can make more of it than I did

              RL

              Working on my Ph.D.  "University of hard Knocks"

              I will consider the opinion that my winnings are a product of chance if you are willing to consider

              they are not.  Many great discoveries come while searching for something else

              USAF https://en.wikipedia.org/wiki/Prime_Base_Engineer_Emergency_Force

                US Flag Trump / 2016 & 2020  

                Avatar
                NASHVILLE, TENN
                United States
                Member #33372
                February 20, 2006
                1044 Posts
                Offline
                Posted: June 29, 2010, 11:59 am - IP Logged

                1 0 0 1 0 0 1 1 0 1   = 589

                1 1 0 1 0 0 1 0 0 0   = 840

                0 0 1 1 0 0 0 1 0 1   = 197

                1 0 0 1 0 0 0 0 1 1   =  579

                0 0 1 0 1 1 0 1 1 0   =  182

                 

                RL

                Another method you might try is to see all those zero's and one's as a binary number.  I took the liberty of calculating the decimal equivalent of each row of binary numbers.

                You could also track the columns as binary number.

                 

                  1    0    0    1    0    0    1    1    0    1   = 589

                  1    1    0    1    0    0    1    0    0    0   = 840

                  0    0    1    1    0    0    0    1    0    1   = 197

                  1    0    0    1    0    0    0    0    1    1   =  579

                  0    0    1    0    1    1    0    1    1    0   =  182

                ________________________________

                 26   8    5    30   1    1  24  21   3    22

                Comparing the columns (not rows) to the next drawing might yield some worthy results.  Let's look at that first column.  Taking only the last 5 results yielded a decimal sum of 26.  What was the next drawing?  Let us imagine the result was "1".  Would a result of 26 always yield future "1"?  How about most of  the time? Or almost never?  50% of the time?  You will have to backtrack to find out.

                Why use the last 5 drawing results?  Why not the last 16? or 20?  Woiuld increasing the database result in more accurate data?  You will have to backtrack to find out.

                Now the next step forward is nothing but quicksand.


                  Canada
                  Member #87735
                  March 2, 2010
                  10 Posts
                  Offline
                  Posted: June 29, 2010, 8:40 pm - IP Logged
                  Can you explain how you calculate decimal sum on binary ?
                    themagician's avatar - artificial intelligence-1.jpg

                    France
                    Member #92435
                    June 7, 2010
                    43 Posts
                    Offline
                    Posted: July 1, 2010, 8:14 am - IP Logged
                    Can you explain how you calculate decimal sum on binary ?

                    Hi,

                    I think he is not taken in consideration the lowest value in the colum to calculate the sum, and so 26 is: 1 1 0 1 in binary, I think that that's how he does it.

                    It is another way to find patterns, you need to test on it, it is only a matter of continue to experiment on it,

                    ;-)

                    Imagination can take you anywhere

                    http://www.artificialintelligencelotto.com 

                      Avatar
                      NASHVILLE, TENN
                      United States
                      Member #33372
                      February 20, 2006
                      1044 Posts
                      Offline
                      Posted: July 1, 2010, 12:38 pm - IP Logged

                      If you are using Windows, you have a calculator which will convert decimal into binary and vise-versa.  Click on "Start", then "All Programs", then "Accessories", and find "Calculator".

                      On the left hand side you will see "HEX", "DEC", "OCT",  and "BIN".  Click on "DEC", input a decimal number, then click on "BIN".  The transition is made.

                      To go from BINary to DECimal, click on BIN, input your binary number, then click on DEC.  The transition is made.

                        Avatar
                        NASHVILLE, TENN
                        United States
                        Member #33372
                        February 20, 2006
                        1044 Posts
                        Offline
                        Posted: July 2, 2010, 4:38 pm - IP Logged

                        Hi,

                        I think he is not taken in consideration the lowest value in the colum to calculate the sum, and so 26 is: 1 1 0 1 in binary, I think that that's how he does it.

                        It is another way to find patterns, you need to test on it, it is only a matter of continue to experiment on it,

                        ;-)

                        Correct me if I am wrong but 1101 in binary would be 13.


                          United States
                          Member #81843
                          October 31, 2009
                          856 Posts
                          Offline
                          Posted: July 6, 2010, 2:40 am - IP Logged

                          Neuroxl has an add-on for excel. Careful tho, they are in the old russia and have a lousy distributer last time I looked. Many sites for the home company (and names?) and one I think was In the Maldives.

                          The site that came up on a G search shows a big sale going on. Keep me informed and PM me if you find anything out about this stuff. It looked good and better than most I found or had, just never took the leap. Make a purchase with a visa gift card to limit any bad biz from the distributer, they have a history.

                          DD

                          Nice signature RL and Ditto on randomness. The creation was not made in disorder and chaos. GEN 1:1 created is perfection implied.

                            themagician's avatar - artificial intelligence-1.jpg

                            France
                            Member #92435
                            June 7, 2010
                            43 Posts
                            Offline
                            Posted: July 6, 2010, 11:32 am - IP Logged

                            Correct me if I am wrong but 1101 in binary would be 13.

                            Sorry, I was wrong 11010 is 26 and he is taking numbers in columns

                            Imagination can take you anywhere

                            http://www.artificialintelligencelotto.com 

                              themagician's avatar - artificial intelligence-1.jpg

                              France
                              Member #92435
                              June 7, 2010
                              43 Posts
                              Offline
                              Posted: July 6, 2010, 12:05 pm - IP Logged

                              I was taking a look at Neurolox, very interesting, it has a predictor add-on, for example to predict the result of a dog race it take one dog past data: dog's weight, the time of the day, and for the past output it takes the dog's timing in past races.
                              As always for artificial intelligence all is a matter of parameters,
                              but how could I pick up the parameters for a Pick 5 data past drawings?
                              Let's take number 15 for example, maybe I could take the total number of times it came out, the number of times it skipped, the position by frequency, but then what could be the output?
                              Past output data is needed for the training,
                              And then I need to take the actual output for every number and compare all the numbers to try to find the "winners"
                              Uh, I really need to think about this one ;)

                              Imagination can take you anywhere

                              http://www.artificialintelligencelotto.com