Sample Questions
Data Science
Directions: Choose the correct option.
Q1. A 4-input neuron has weights of 3, 4, 2, and 1. The activation function is linear, with the
constant of proportionality equal to 2. A bias of 10 exists at the nodes. The inputs are 4, 10, 20,
and 5, respectively. What is the output of node?
A. 204
B. 194
C. 184
D. 107
Directions: Choose the correct option.
Q2. A realtor wishes to predict house prices in an area. She decides to use the following
variables: distance from one's office, distance from the community park and distance from a
public school. You can assume they are independent of each other. She builds a linear
regression model on this data to predict the prices. The following are the coefficients of the
trained model and some additional parameters: coefficients, standard error of the coefficients,
t and p values:
Based on this information, which of these variables should she NOT use in building her model?
A. Distance from the office
B. Distance from the community park
C. Distance from the public school
D. Cannot decide based on the information given
© Aspiring Minds Assessment Pvt. Ltd.
Directions: Choose the correct option.
Q3. Interquartile range is defined as the distance between the 75th percentile and the 25th
percentile. Given a one-dimensional data set, which of the following statistics will be influenced
the most by the outliers in the data set?
a. Interquartile range
b. Median
c. Mode
d. Range
A. Only d
B. a and b
C. a and d
D. a, b, c and d
Directions: Choose the correct option.
Q4. You decide to build a ridge regression model to solve a problem. You find that your data
set has many features that are highly correlated, and you decide to use principal component
analysis (PCA) for dimensionality reduction. You also need to use cross-validation for the
parameter(λ) selection in ridge regression. What are the proper steps for this model-building
task?
A. Divide the data into training and test sets. Use PCA on the training data to create a new set
of data that is of fewer dimensions. Apply cross-validation on the transformed training data
by partitioning it into training and validation sets to select λ. Apply PCA on the test data to
transform them into data with similar dimensions as the transformed training data. Assess
the model using transformed test data.
B. Use PCA to create a new set of data that is of fewer dimensions. Divide the data into
training and test sets. Apply cross-validation on the training data by partitioning it into
training and validation sets to select λ. Assess performance on the test data.
C. Divide the data into training and test sets. Use PCA on the training data to create a new
set of data that is of fewer dimensions. Apply cross-validation on the transformed training
data by partitioning it into training and validation sets to select λ. Project the test data onto
the same low-dimension space as that of the transformed training data. Assess the model
using the transformed test data.
D. Divide the data into training and test sets. Partition the training data into training and
validation sets and use cross-validation to select λ. Use PCA on the training data to create a
new set of data that is of fewer dimensions. Make a ridge regression model using this low-
dimension training data. Project the test data onto the same low-dimension space as that of
the transformed training data. Assess the model using the transformed test data.
© Aspiring Minds Assessment Pvt. Ltd.