18.3: sklearn

Scikit-learn, often abbreviated as sklearn, is a versatile and widely-used open-source machine learning library for Python. It provides a range of supervised and unsupervised learning algorithms through a consistent interface. Essential tools for data mining and data analysis are also part of this library, making it a go-to choice for many data scientists and researchers. Scikit-learn is built upon the SciPy (Scientific Python) ecosystem and is designed to interoperate with the Python numerical and scientific libraries NumPy and SciPy. Its simplicity, accessibility, and broad range of algorithms make it very useful for educational purposes, yet it's powerful enough to be used in large-scale industrial applications. Scikit-learn includes a variety of algorithms for classification, regression, clustering, dimensionality reduction, model selection, and preprocessing. Its robustness, ease of use, and good documentation have made it one of the most popular machine learning libraries in the world.

While sklearn can do a lot of sophisticated analyses, it can also do the simplest, a linear fit.  To do so, follow these steps:

1. Import the necessary classes: Use the `LinearRegression` class from the `sklearn.linear_model` module.

2. Prepare your data:
   - Ensure your independent variable(s) (features) are in a two-dimensional array (or matrix). If you have a single feature, you'll need to reshape your array to be two-dimensional.
   - The dependent variable (target) should be in a one-dimensional array.

3. Create an instance of the LinearRegression class:
   - Instantiate the `LinearRegression` class to create the model.

4. Fit the model:
   - Use the `fit` method on the model instance with your independent and dependent variables as arguments. This step is where the model learns the coefficients from your data.

5. Check the coefficients:
   - After fitting the model, you can check the `coef_` and `intercept_` attributes of your `LinearRegression` object to see the slope and intercept of the fitted line.

6. Make predictions:
   - With the model fitted, you can now predict the dependent variable for new independent variable(s) using the `predict` method.

Here's an example of how to do a linear fit with scikit-learn for a dataset with a single feature:

from sklearn.linear_model import LinearRegression
import numpy as np

# Sample data
X = np.array([[1], [2], [3], [4], [5]])  # Independent variable (feature), needs to be 2D for sklearn
y = np.array([2, 3, 5, 7, 11])           # Dependent variable (target), 1D is fine

# Create an instance of the model
model = LinearRegression()

# Fit the model
model.fit(X, y)

# After fitting the model, we can check the slope (coef_) and intercept (intercept_)
slope = model.coef_[0]
intercept = model.intercept_

# Make predictions (optional)
X_new = np.array([[6], [7]])  # New data for prediction
y_pred = model.predict(X_new)

# slope, intercept, and predictions are now available
print(f"Slope: {slope}, Intercept: {intercept}, Predictions: {y_pred}")

This code will output the slope and intercept of the best fit line, as well as predictions for the `X_new` data points. If you have multiple features, `X` would be a matrix with multiple columns, one for each feature. The `y` array remains one-dimensional.