Python Qt6 : testing CatBoost with loto 6/49 numbers.

CatBoost is a high-performance library for gradient boosting on decision trees, developed by Yandex and used for various tasks. It supports categorical features, GPU version, improved

See the official webpage.

python -m pip install catboost
Collecting catboost
...
Installing collected packages: narwhals, plotly, catboost
Successfully installed catboost-1.2.8 narwhals-2.15.0 plotly-6.5.2

The source code implements a single, clean, fully self‑contained Python application that relies exclusively on real lottery data loaded from a text file. It computes all required numerical relationships for each draw, including digit distribution, statistical measures, and parity-based metrics. Using only the real historical draws, the program automatically builds a training dataset where each row uses the previous extraction as input and the next extraction’s digit distribution as the prediction target. It then trains two CatBoost regression models—one for forecasting how many one‑digit numbers will appear in the next draw, and another for predicting how many two‑digit numbers will appear. All computed metrics and model predictions are displayed in a structured PyQt6 table, ensuring that every result is derived entirely from the real data provided by the user.

I used 217 old real numbers from loto 6/49:

1 6 12 16 17 33
17 26 30 35 36 44
23 24 28 34 48 49
5 11 22 34 42 43
7 12 17 30 33 49
12 23 32 41 43 48
4 6 33 35 36 39
4 6 33 35 36 39
13 14 20 21 38 49
4 9 11 15 37 47
...

This is the result: