Jeremiahcp

On 6/14/2019 at 7:54 AM, morganstark said:

Please forgive me for any grammar mistake.

I'm a civil engineer and completed my Msc (Maths) focusing on Numerical Study 10 years ago. After my semi retirement as a result of my financial freedom, i have been studying some practical Maths problem for fun.

Recently I've been trying to model and solve a 2 digit lottery drawing game, and i failed. It's purely my imagination since i didn't see this in anywhere. But who knows it may exist? 

Suppose we have a lottery game of 2 digits, drawn from 2 separate but identical electrical drums as lottery company always have. Each drum consists of 10 balls, numbered from 0 to 9, to be drawn as a pair and the drawn balls are to be replaced. In one game, 12 pairs of numbers to be drawn as winning numbers, on every Saturday and Sunday. 

Eg
A particular Saturday: 09, 21, 04, 31, 48, 61, 00, 32, 99, 98, 11, 99
Sunday: another 12 pairs of numbers

My question is: if you have the result of last 1000 game, how do you calculate the most probable drawn numbers (one or two pairs)  for the next drawing? 

Any idea? 

Make a histogram and check the relative frequency. If you have something that looks like a box, then there is equal probability, if not, then you can see which ones pop up more often in the long run. Assuming a 1000 games is enough to give you an accurate picture, most lotteries will have such a span of possible combinations 1000 games may not be good enough.  So you could run some computer simulations to see just how well this approach works with your set up. 

I think you would need to be more explicit in what you want to learn. If it is statistics I can make recommendations, but I have very little idea how the other sides do things. 

However, I think generally you'd want to make sure your algebra and calculus is top notch, then add in some basic understanding of proofs and linear algebra. At the very least. 

R^2 is not a measure of correlation. It is a measure of variability in the dependent  variable explained by changes in the independent variables. For example if you have an R^2 of 24% that means your regression model explains 24% of the variability in the dependent variable, which means you are missing 76% of the needed information. 

Now you can take the square root of R^2 to get Pearson's r, and that is an estimate of the correlation. But keep in mind that correlation is a very specific type of association, one that is linear in nature. However, you can still have non linear associations, at which point you would either need to add a ploy term to your model, do a transformation or use more advanced options. 

A low R^2 means your model is a poor fit, and a low p-value does not mean  you have a working model, it could merely mean that what you have is better at fitting the response than nothing at all. You would want to do ANOVA, AIC, BIC, or some other type of model comparison to see if you model works better than a mean only model. 

If you want to build a regression model use model selection steps to find a better fitting model, if you merely want to report on the correlation between two variables regardless of whatever else is going on, you can use Person's r as a statistical summary for that measure of correlation. A decent fitting regression model would allow you to explore variable correlation in much more depth, but if your model does not meet the assumptions for a linear model and is a poor fit you will not get reliable information.