classif1.py

classif1.py

%matplotlib qt
from sklearn.dummy import DummyClassifier
import pandas as pd
import seaborn as sns

# Load the penguins dataset from seaborn
penguins_df = sns.load_dataset("penguins")

# Split the dataset into features (X) and target variable (y)
X = penguins_df[['bill_length_mm', 'bill_depth_mm', 'flipper_length_mm', 'body_mass_g']]
y = penguins_df['species']

# review quickly the populations
print(y.value_counts(normalize=True))

def plot_dummy_strategy_classifier(
        X, y, target_name="species", 
        strategies=['most_frequent', 'prior', 'stratified', 'uniform', "constant"],
    ):
    # Note that we use stratify=True to keep the class distribution both in the
    # train set and test set
    X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y)
    dfs = []
    dfs_proba = []
    df_true = pd.DataFrame(pd.Series(y_test, name="species"))
    df_true['strategy']="ground_truth"
    dfs.append(df_true)
    for i, strategy in enumerate(strategies):
        dummy = DummyClassifier(strategy=strategy, constant="Chinstrap")
        dummy.fit(X_train, y_train)
        df = pd.DataFrame(pd.Series(dummy.predict(X_test), name=target_name))
        df['strategy']=strategy
        dfs.append(df)

        df = pd.DataFrame(dummy.predict_proba(X_test), columns=["Adelie", "Chinstrap", "Gentoo"])
        df['strategy']=strategy
        dfs_proba.append(df)

    df = pd.concat(dfs)
    df_proba = pd.concat(dfs_proba)
    return df, df_proba

df, df_proba = plot_dummy_strategy_classifier(X, y)

sns.catplot(
    data=df, kind='count', x="species", col='strategy', hue="species", 
    hue_order=["Adelie", "Gentoo", "Chinstrap"], order=["Adelie", "Gentoo", "Chinstrap"], # so the order is reproducible
)

fig, axes = plt.subplots(1, df_proba['strategy'].nunique())
cbar_ax = fig.add_axes([0.91, 0.15, 0.02, 0.7])  # Adjust these values to position the colorbar as desired
for ax, strat in zip(axes, df_proba['strategy'].unique()):
    sns.heatmap(df_proba.query('strategy==@strat')[["Adelie", "Gentoo", "Chinstrap"]], ax=ax, cbar=False, vmin=0, vmax=1, yticklabels=False)
    ax.set_title(f'strategy = {strat}')
fig.colorbar(ax.collections[0], cax=cbar_ax)
sns.displot(df_proba, col="strategy", hue="species"