Wine 🍷 Dataset Exploration and XGBoost Regression Hyperparameter Tuning with Optuna & AutoML

Classic Take on the Wine Review Dataset from Kaggle

Part 1 - EDA Exploration with dataprep

Objective: Understand data distribution -> Python_file_EDA// Notebook_EDA

Part 2 - Data Extraction / Feature Eng / Data Encoding

Objective: Extract features from text fields, One Hot Encoding of Categorical Features ->> Data_Processing

Part 3 - Train XGBoost to Predict Wine Score, Parameter Hypertuning with Optuna, Feature Importance

Objectives: Train model, tune with bayesian hyperparameter optimization (Optuna), Evaluate feature importance -> Notebook // Python_file

• A) Create XGBoost model to predict Wine score based on Wine Origin, Price and description features.

• B) Use Optuna to tune hyperparameters, evaluate the most important hyperparamenters.

• C) Based on tuned parameters evaluate model feature importance.

Part 4 - AutoMl with mljar-supervised

Objective: Run AutoMl on same Wine dataset, compare performance ->Notebook

• A) AutoMl as model EDA and Explained -> Explainer
• B) AutoML as a powerful model/hyperparameter tuner ->Model_tuner // Optuna-AutoML-XGBoost