Feature scaling (StandardScaler, MinMaxScaler) in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why use StandardScaler instead of MinMaxScaler?

Imagine you have a dataset with features measured in very different units, like height in centimeters and income in dollars. You want to prepare the data for a machine learning model.

Why might you choose StandardScaler over MinMaxScaler?

A StandardScaler centers data to mean 0 and scales to unit variance, which handles outliers better than MinMaxScaler.

B StandardScaler always scales data between 0 and 1, which is better for all models.

C MinMaxScaler removes outliers automatically, so StandardScaler is better for noisy data.

D MinMaxScaler centers data to mean 0 and scales to unit variance, so StandardScaler is not needed.

❓ Predict Output

intermediate

Output of scaling with MinMaxScaler

What is the output of the following code?

ML Python

from sklearn.preprocessing import MinMaxScaler
import numpy as np

X = np.array([[1], [2], [3], [4], [5]])
scaler = MinMaxScaler()
X_scaled = scaler.fit_transform(X)
print(X_scaled.flatten())

A [0. 0.25 0.5 0.75 1. ]

B [1. 2. 3. 4. 5.]

C [-1. -0.5 0. 0.5 1.]

D [0.2 0.4 0.6 0.8 1. ]

❓ Metrics

advanced

Effect of StandardScaler on mean and variance

After applying StandardScaler to a feature column, what will be the approximate mean and variance of the transformed data?

A Mean close to 1 and variance close to 0

B Mean close to 0 and variance close to 1

C Mean close to 0 and variance close to 0

D Mean close to 1 and variance close to 1

🔧 Debug

advanced

Why does this scaling code raise an error?

Consider this code snippet:

from sklearn.preprocessing import StandardScaler
import numpy as np

X = np.array([1, 2, 3, 4, 5])
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)
print(X_scaled)

Why does this code raise an error?

A fit_transform method does not exist for StandardScaler, causing AttributeError.

B StandardScaler cannot handle integer data types, causing a TypeError.

C Missing import for StandardScaler causes NameError.

D X is 1D array but fit_transform expects 2D array, causing a ValueError.

❓ Model Choice

expert

Choosing scaler for neural network input

You are training a neural network on image pixel data with values from 0 to 255. Which scaler is the best choice to preprocess the input pixels before training?

A No scaling needed because neural networks handle raw pixel values well

B StandardScaler to center pixels around 0 with unit variance

C MinMaxScaler to scale pixel values between 0 and 1

D Use a scaler that scales pixels between -1 and 1 by subtracting 128 and dividing by 128