Python Fit Transform

In machine learning, fit (), transform(), fit transform () methods are provided by scikit learn package. This package is used in model fitting and data processing. These methods are implemented using in-built functions in this package.

• The fit() method is used to determine the mean and standard deviation for a given aspect to be used for scaling.
• The transform() is used to implement scaling using mean and standard deviationevaluated using the.fit() method.
• The fittransform() method has bothtransform and fits.

Packages required for solving problems:

• pip install pandas
• pip install scikit-learn

Example 1:

// python program
fromsklearn importdatasets
importpandas as x
  
iris =datasets.load_iris()
data =x.DataFrame(iris.get('data'), columns=[
    'sepal length', 'petal length', 'sepal width', 'sepal width'])
data.head()

Output:

The fit () Method

The fit function determines the statement to transform the column converts the column of interest by transforming but it doesn’t allot original transformation. It calculated is stored as a fit object. It doesn’t return any value.

Example:

fromsklearn.preprocessing importStandardScaler
  
scaler =StandardScaler()
scaler.fit(data['sepal width'])

Output:

standardScalar ()

The transform () method:

The transform () takes superiority of the fit obj in the fit() and allocates the existed transformation onto the column. So, fit() and transform() is a two-step process that finishes transformation in the next step. Here transform returns the existed transformed array but fit () method doesn’t return anything.

Example:

scaler.transform(data['sepal width'])

Output:

Fit Transform () Method

In transform () and fit () has two-step process, so we use fit transform () for single shot process. In fit transform () for calculating and apply the transformation in a one step.

Example:

scaler.fit_transform(X_train)

Output:

As we observe, fit (), transform (), fit transform () has same output. These methods use same transformation is assigned to test dataset. But fit () doesn’t use test dataset. So, we use transform () directly on test dataset.

Example:

scaler.transform(X_test)

Output: