# #1 Solution 

Sorry for the late post.
Tried a few approaches but found that light gradient boosted models (lightgbm) worked the best.
In this solution, I show lightgbm modeling over full training data set as lightgbm tends to do well even with low information data.
Data was trained using a 80/20 training/validation split to find optimal parameters for model. The parameters was optimized utilizing over 1000 iterations of random selection of 6 different parameters
Afterwards, model was trained on full data set to predict on test data set
code can be found
[here](https://www.kaggle.com/code/paulyy27/tfug-diabetes-prediction-using-lightgbm/)