tree decision model by EJARA WORKU

[ejara42]

EDA: shows structure, missing values, basic stats so you understand the dataset.

Target selection: automatically chooses a target column, but you can override.

Preprocessing: numeric imputation (median), categorical imputation + one-hot encoding.

Train/test split: hold out 20% for final evaluation; stratify if classification.

Baseline model: Decision Tree pipeline trained with default params.

Evaluation: accuracy/MSE, confusion matrix, classification report.

Cross-validation: quick 5-fold check.

Hyperparameter search: GridSearchCV to improve the tree.

Visualization: plot top levels of the tree and print important features.

Save: persist model and evaluation to model_artifacts, and optionally copy to Drive