Coding exercises and experiments
The task involved analyzing wine data using sklearn.datasets.load_wine. It required a preliminary analysis, building models with XGBoost and DNN, and training and validating results. The impact of parameter and architecture changes on model performance was analyzed and compared.
Note: Tasks completed in Polish.
The task focused on preparing a dataset with missing values to train a classifier predicting heart disease (num as target). This included handling missing data, optimizing learning rate and steps with xgb.cv, and avoiding overfitting. The final model was trained based on error trends and tree count optimization.
Note: Tasks completed in Polish.
Task 1: A script was created to generate a temperature report for a given day, converting sensor values and displaying results with precision and a visual bar representation. Bonus: Added validation for valid day input.
Task 2: Using an NBA dataset, tasks included identifying non-USA players who attended Kentucky, analyzing height trends, identifying top scorers, and assessing draft round distributions for players over 100 kg.
Task 3: Implemented a create_tower function that generates a pyramid-like structure based on input size.
Task 4 (Bonus): Developed a "Tic-Tac-Toe" game with real-time board updates, turn-based inputs, and automatic win/draw detection.
Note: All tasks completed in Polish.