Neural regression checkpoint · Data Playground

Retrieval 1 · Wine600

Complete the Wine600 neural regression workflow with feature/target scaling, width selection, convergence evidence and original-unit errors.

This practice: Complete the Wine600 neural regression workflow with feature/target scaling, width selection, convergence evidence and original-unit errors. Next: Clustering and Discovery. Continue to Clustering and Discovery →

Given data · Wine600

600 observations. One red or white wine sample. The dataframe df is supplied afresh for each Run.

Download source CSV · Source and original dictionary

Wine600 · first 8 prepared rows
fixed acidity	volatile acidity	citric acid	residual sugar	chlorides	free sulfur dioxide	total sulfur dioxide	density	pH	sulphates	alcohol	quality	wine_type
5.3	0.32	0.12	6.6	0.043	22	141	0.9937	3.36	0.6	10.4	6	white
6.4	0.3	0.27	4.4	0.055	17	135	0.9925	3.23	0.44	12.2	6	white
6.5	0.22	0.32	2.2	0.028	36	92	0.99076	3.27	0.59	11.9	7	white
8	0.24	0.48	6.8	0.047	13	134	0.99616	3.23	0.7	10	5	white
8.9	0.29	0.35	1.9	0.067	25	57	0.997	3.18	1.36	10.3	6	red
6.6	0.21	0.39	2.3	0.041	31	102	0.99221	3.22	0.58	10.9	7	white
8.8	0.24	0.54	2.5	0.083	25	57	0.9983	3.39	0.54	9.2	5	red
5.4	0.15	0.32	2.5	0.037	10	51	0.98878	3.04	0.58	12.6	6	white

Column meanings and units

quality: ordered sensory score (0–10). alcohol: volume percent. pH: acidity scale. Other chemistry units follow the linked source dictionary.

The score is ordinal but modelled as regression here. ML removes exact duplicate rows before splitting. Chemistry must be available at prediction time; predictive associations do not establish effects of changing an ingredient.

Input schema
Column	Stored type
fixed acidity	float64
volatile acidity	float64
citric acid	float64
residual sugar	float64
chlorides	float64
free sulfur dioxide	float64
total sulfur dioxide	float64
density	float64
pH	float64
sulphates	float64
alcohol	float64
quality	int64
wine_type	str

Declared validation design

Reserve 20% of the declared population for the final test using split seed 42. Use a random split.

Use the same five training folds, shuffled with seed 42, for reference comparison, candidates and selection. Use out-of-fold training predictions for diagnosis. Select with negative RMSE (larger is better) and compare with a training-mean reference. Report final RMSE in original target units. Open final-test evidence after selection and diagnosis.

Supporting concepts: Neural regression workflow → · Convergence, early stopping and validation →

Remember the idea

This checkpoint combines previously taught skills. Assemble the workflow; help remains available when needed.

Schematic · Complete the Wine600 neural regression workflow with feature/target scaling, width selection, convergence evidence and original-unit errors.Scroll the diagram horizontally if needed.

Required Python variables and evidence

Use these names so Check can inspect your workflow. Each meaning is shown beside its name.

Workflow evidence contract
Variable	Meaning
X	Feature dataframe for the declared population, preserving row indices.
X_test	Final-test feature rows from the declared split.
X_train	Training feature rows from the declared split.
cv_results	Candidate validation evidence; comparison workflows use a dataframe indexed by model ID.
final_model	Chosen pipeline fitted on training rows, after selection and diagnosis.
final_predictions	Unaltered predictions of final_model on X_test.
final_rmse	Root mean squared error in original target units.
loss_curve	Training loss history of the selected neural regressor.
reference_results	cross_validate result for the dummy reference; test_score contains five scores.
residuals	Training-only actual, predicted and actual-minus-predicted residual evidence.
search	Fitted GridSearchCV for the declared parameter comparison.
y	Target series, aligned with X.
y_test	Final-test targets, aligned with X_test.
y_train	Training targets, aligned with X_train.

Hint 1 — Think

Reconstruct both feature and target preparation inside every neural fit.

Hint 2 — Tools

Pipeline, TransformedTargetRegressor, nested width search, loss curve and original-unit RMSE.

Hint 3 — Approach

Compare against the mean reference, select width on training folds, inspect convergence and residuals, then predict through the fitted wrapper.

Explained solution

from sklearn.base import clone
from sklearn.model_selection import train_test_split, KFold, StratifiedKFold, cross_validate, cross_val_predict, GridSearchCV
from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer, TransformedTargetRegressor
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.dummy import DummyRegressor, DummyClassifier
from sklearn.metrics import root_mean_squared_error, f1_score, accuracy_score, confusion_matrix

from sklearn.neural_network import MLPRegressor
X = df.drop(columns=['quality', 'wine_type'])
y = df['quality']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
model = Pipeline([('prepare', StandardScaler()), ('model', TransformedTargetRegressor(regressor=MLPRegressor(hidden_layer_sizes=(24,), max_iter=800, early_stopping=True, tol=1e-3, random_state=42), transformer=StandardScaler()))])
folds = KFold(n_splits=5, shuffle=True, random_state=42)
cv_results = cross_validate(model, X_train, y_train, cv=folds, scoring='neg_root_mean_squared_error')
reference_results = cross_validate(DummyRegressor(strategy='mean'), X_train, y_train, cv=folds, scoring='neg_root_mean_squared_error')
search = GridSearchCV(model, {'model__regressor__hidden_layer_sizes':[(16,), (24,)]}, cv=folds, scoring='neg_root_mean_squared_error')
search.fit(X_train, y_train)
oof_predictions = cross_val_predict(search.best_estimator_, X_train, y_train, cv=folds)
residuals = y_train - oof_predictions
loss_curve = search.best_estimator_.named_steps['model'].regressor_.loss_curve_
final_model = clone(search.best_estimator_).fit(X_train, y_train)
final_predictions = final_model.predict(X_test)
final_rmse = root_mean_squared_error(y_test, final_predictions)
print('Final RMSE in quality-score units:', final_rmse)

Fold-local target scaling supports optimisation without leakage, and the wrapper reverses it so final errors retain quality-score units.

Helpful prior knowledge: Neural regression workflow · Shared network retrieval These links are guidance, not locks.

Sources and API context

Examples run with this Playground’s scikit-learn 1.4.2 / Pyodide 0.26.4 runtime.