Saving and Loading Machine Learning Models

Before deploying a model, you need to save it to disk so you don't have to retrain it every time you use it.

Step 1: Train the Model

from sklearn.linear_model import LogisticRegression

model = LogisticRegression()
model.fit(X_train, y_train)

Step 2: Save the Trained Model

The joblib library is commonly used to save Scikit-learn models efficiently, especially when the model contains large numerical data.

import joblib

joblib.dump(model, "model.joblib")

Step 3: Load the Saved Model

Once saved, you can load the model back at any time, without retraining, using the same library.

loaded_model = joblib.load("model.joblib")

Step 4: Use the Loaded Model for Predictions

predictions = loaded_model.predict(X_test)
print(predictions)

You can now move this saved file to a server, an application, or any environment where the model needs to run, making it the foundation for deployment.