Polanitz/Forecasting-the-Next-Winning-Numbers-in-the-Texas-Lottery-Mega-Millions-Drawing-using-A-Deep-Neura ? reverse-engineered prompt

Build me a simple, self contained lottery forecasting project around the Texas Mega Millions historical CSV in this repo. I want it to use a neural network in a notebook, probably an LSTM style time series approach, to learn from past drawings and then give a best guess for the next set of numbers. Keep it practical and easy to follow, since this is more of a fun experiment than a serious gambling tool.

Please load and clean the CSV, turn the past drawings into rolling windows, train the model, and show the results clearly with a few charts and basic evaluation so I can see whether it learned anything useful at all. At the end, print a predicted next draw in a readable format, and include a short note about the limits of this kind of prediction. Add comments in plain English throughout so a beginner can understand what each step is doing. If anything is unclear from the data format, look up current docs online if you need to.