SciKit-Learn

What is SciKit-Learn?

SciKit-Learn is a [Python][-py] library that offers support for implementing the vast majority of machine learning algorithms. It is arguable the most useful ML library in Python because it contains many functions and tools for implementing predictive algorithms and performing statistical analysis of data.

Install

The scikit-learn library can be installed using pip. To install, just enter the following command in your terminal:

pip install scikit-learn

Once installed, you can verify it works by opening the Python REPL, using the python command by itself. Then trying to import the library by entering the following command:

import sklearn

Training vs Testing Datasets

Before diving deeper into how to use SciKit-Learn, it is important to understand how to treat the data that is provided to make the desired prediction.

Suppose you are given a dataset with 1000 rows and two columns, one that contains the hours that a student spent on an assignment and the other contains the grade on that assignment. Assume that you want to use linear regression to predict the grade of a student based on how many hours were spent on the assignment.

Given that you are provided with only one dataset, you need to find a way to use the data both to train the linear regression algorithm to make an accurate prediction and to test the predictions. In other words, you want to split the data into a training set and a testing set. The training set will be used to train the predictive algorithm, and the testing set will be used to test the accuracy of the algorithm.

Consider the following example. Suppose you are given a dataframe, df, that contains the two columns describe above. The first step will be to seperate the data into two dataframes, X and y. The X is the conventional name for the dataframe that contains feature data that are independent that will be used to make the prediction of the dependent or target variable, y.

Then, you can define X and y to be:

X = dataset.iloc[:, :-1].values
y = dataset.iloc[:, 1].values

After this step, you need to split each set into a training and a testing set. This can be quickly accomplished by using the train_test_split function in the sklearn.model_selection module.

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y)

Here the code splits X and y into training and testing sets. The testing set will contain 20% of the original data. Setting the random_state variable equal to a certain value (in this case 0) will ensure reproducibility of the results.

Note: This is a fundamental step in machine learning. Using too much of the data to train the model will risk overfitting the model. This makes future predictions more brittle and less responsive to real-world data.

Select Applications of SciKit-Learn

SciKit-Learn is extremely popular, not only because it offers access to a wide range of ML algorithms, but also because it was built so almost any learning algorithm can be built using it.

First Example Application

This example illustrates how to implement linear regression using SciKit-Learn. The first step is to import the necessary libraries. Because you want to implement linear regression, you need to import the linear_model module from SciKit-Learn.

from sklearn import linear_model,train_test_split

Next suppose that you have a dataframe, df. As explained, you need to define which columns of the df dataframe will be used as X values and which will define the variable target or dependent variable, y.

X = df.iloc[:, :-1].values
y = df.iloc[:, 1].values

Then, you need to split the data into training and testing sets.

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=0)

After doing so, you are ready to train your linear regression algorithm. First, initialize the r classifier:

r = LinearRegression()

Once this is done, you can retrieve the intercept and the coefficient of the line of best fit for your data:

print(r.intercept_)
print(r.coef_)

Finally, you can use the predict method to make predictions on the testing set:

y_pred = r.predict(X_test)

Although the example above applies to linear regression, the fundamental idea to keep in mind is that essentially any ML algorithm can be used in SciKit-Learn by following these same steps.

A comprehensive list of examples can be found in the SciKit-Learn Docs.

References

Web Links

Note Links