The IIT Madras BS Machine Learning Practice (MLP) End Term paper sat on 28 Apr 2024, in the January 2024 term: 39 questions for 100 marks in 180 minutes. Every question is below with its answer. Take it as a timed mock test to be marked, or read it through first.
Select multilabel multiclass classification problems:
There is a collection of photographs. Each photograph can have multiple animals, e.g., cats, dogs and birds. Your model should indicate all the animals which are present.
From appropriate weather data, your model must predict, average temperature and average humidity for next seven days.
Predicting expected price of a second hand car with appropriate features.
None of these.
Correct answer
There is a collection of photographs. Each photograph can have multiple animals, e.g., cats, dogs and birds. Your model should indicate all the animals which are present.
Correct answer
Which of the following can be used (with appropriate supporting code) to compute training error after each iteration?
Correct answer
For the given code below in which you use how many models will get trained or what will be the length of scores variable ?
from sklearn.model_selection import cross_val_scorefrom sklearn.model_selection import LeaveOneOutfrom sklearn.linear_model import LogisticRegressionfrom sklearn.datasets import make_classificationX, y = make_classification(n_samples=1024, n_features=82, n_classes=2,↪ random_state=42)estimator = LogisticRegression()loocv = LeaveOneOut()scores = cross_val_score(estimator, X, y, cv=loocv)1106
82
1024
5
None of these
Correct answer
1024
What is the purpose of k-fold cross-validation ?
To split data into training and testing sets.
To tune hyperparameters.
To evaluate model performance on multiple subsets.
To preprocess data.
Correct answer
To evaluate model performance on multiple subsets.
Statement 1 is True and Statement 2 is False
Statement 2 is True and Statement 1 is False
Both the statements are True
Both the statements are False
Correct answer
Statement 1 is True and Statement 2 is False
Which assumption does Naive Bayes make about the features?
They are independent of each other.
They are always a numerical representation of the text data.
They are linearly related.
They are categorical.
None of these.
Correct answer
They are independent of each other.
Given the following code snippet that preprocesses a dataset with both continuous and categorical features using sklearn.preprocessing tools, what will be the first row of the X_transformed array after preprocessing?
import numpy as npfrom sklearn.preprocessing import OneHotEncoderfrom sklearn.preprocessing import MinMaxScaler, StandardScalerfrom sklearn.compose import ColumnTransformer
X = np.array([[4.0, 'avocado'], [3.0, 'dragon fruit'], [2.0, 'sapodilla'], [7.0, 'papaya']])
preprocessor = ColumnTransformer( transformers=[('num1', MinMaxScaler(), [0]), ('cat', OneHotEncoder(), [1]), ('num2', StandardScaler(), [0]),])
X_transformed = preprocessor.fit_transform(X)print(X_transformed[0])Correct answer
You’re working on a dataset containing customer purchase data, and you want to segment the customers into distinct groups based on their purchasing behavior. Each data point represents a customer and includes features like “Total Amount Spent” and “Number of Items Purchased.” Which algorithm is suitable for this scenario?
Linear Regression
Decision Tree
K-means Clustering
Support Vector Machine
None of these
Correct answer
K-means Clustering
Increasing the number of neurons in hidden layers will always lead to better model performance.
Decreasing the number of neurons in hidden layers reduces the model’s capacity to capture complex patterns.
The number of neurons in hidden layers does not significantly affect the model’s performance.
Finding the optimal number of neurons is a trial-and-error process and may require experimentation.
Correct answer
Finding the optimal number of neurons is a trial-and-error process and may require experimentation.
How does agglomerative clustering handle outliers?
It ignores outliers during the clustering process.
It assigns outliers to the nearest cluster.
It creates separate clusters for outliers.
It removes outliers from the dataset before clustering.
Correct answer
It assigns outliers to the nearest cluster.
What is agglomerative clustering?
A hierarchical clustering technique that starts with each data point as its cluster and merges the closest clusters iteratively.
A method for partitioning data into a predefined number of clusters.
A clustering algorithm that uses centroids to iteratively assign data points to clusters.
A dimensionality reduction technique that projects data onto a lower- dimensional space.
Correct answer
A hierarchical clustering technique that starts with each data point as its cluster and merges the closest clusters iteratively.
What initialization method is best in KMeans to select initial cluster centroids?
Random initialization
K-means++ initialization
Hierarchical agglomerative initialization
Weighted initialization
Correct answer
K-means++ initialization
The distance between cluster centroids
The number of clusters formed
Sum of squared distances of samples to their closest cluster center.
The silhouette coefficient
Correct answer
Sum of squared distances of samples to their closest cluster center.
Correct answer
Correct answer
Given the following code using BaggingClassifier with KNeighborsClassifier as the base estimator:
from sklearn.ensemble import BaggingClassifierfrom sklearn.neighbors import KNeighborsClassifier
base_knn = KNeighborsClassifier(n_neighbors=5)
bag_clf = BaggingClassifier(base_knn, n_estimators=50, max_samples=0.5,↪ bootstrap=True, n_jobs=-1)Which of the following statements is correct?
Correct answer
Correct answer
What is the solution for overfitting?
To have less constraints/regularization
To have more constraints/regularization
Delete a significant portion of data
Increase the dataset size.
None of these.
Correct answers
To have more constraints/regularization
Increase the dataset size.
Select all the correct options:
The more the SGD iterations, the lesser the fluctuations in training error.
More iterations require more computation time.
The tol (error tolerance) parameter restricts the number of iterations performed.
Training error might not consistently decrease while performing SGD iterations.
None of these
Correct answers
The more the SGD iterations, the lesser the fluctuations in training error.
More iterations require more computation time.
The tol (error tolerance) parameter restricts the number of iterations performed.
Training error might not consistently decrease while performing SGD iterations.
Correct answers
Which of the following statements are true?
Correct answers
Which of the following options are correct regarding regularization?
It is a technique used to minimize the adjusted loss function and avoid underfitting.
It helps in increasing the bias of the training model.
It determines the rows to be selected as a training dataset.
Elastic net regularization is a combination of L1 and L2 regularization both.
Correct answers
It helps in increasing the bias of the training model.
Elastic net regularization is a combination of L1 and L2 regularization both.
Which of the following techniques are used in decision trees to make decisions or to measure the quality of a split while training the model?
Entropy
RoC Curve
Cross Entropy
Gini Impurity
Correct answers
Entropy
Gini Impurity
Which of the following approaches is(are) helpful to find a good value for k in k−means clustering algorithm?
Plotting an Elbow curve.
Using classifiers before making clusters.
Plotting Silhouette coefficient for various values of k.
Using k-fold cross validation
Correct answers
Plotting an Elbow curve.
Plotting Silhouette coefficient for various values of k.
Consider the following block of code:
from sklearn.datasets import load_breast_cancerfrom sklearn.tree import DecisionTreeClassifierfrom sklearn.model_selection import train_test_splitX,y = load_breast_cancer(as_frame = True, return_X_y = True)X_train,X_test,y_train,y_test = train_test_split(X,y, test_size = 0.2, random_state = 1)clf = DecisionTreeClassifier(min_samples_split = 5, min_samples_leaf = 3, random_state = 5)clf.fit(X_train, y_train)print(clf.score(X_test, y_test))In which of the following scenarios, the split will NOT be made at node N?
10 number of samples at node N. If it is split, it can split such that 2 samples in the left child and 8 samples in the right child.
6 number of samples at node N. If it is split, it can split such that 3 samples in the left child and 3 samples in the right child.
12 number of samples at node N. If it is split, it can split such that 5 samples in the left child and 7 samples in the right child.
4 number of samples at node N. If it is split, it can split such that 3 samples in the left child and 1 samples in the right child.
Correct answers
10 number of samples at node N. If it is split, it can split such that 2 samples in the left child and 8 samples in the right child.
4 number of samples at node N. If it is split, it can split such that 3 samples in the left child and 1 samples in the right child.
Correct answer: 0.5
Consider following data points:
import numpy as npX = np.array([[1,1],[10,11],[5,5],[25,18],[-1,-1]])y = np.array([0,1,0,1,1]).reshape(-1,1)What will be the highest accuracy a perceptron model can achieve on this dataset without any feature engineering?
Correct answer: 0.8
Correct answer: 0.375 (accepted within ±0.005)
Consider the following Python code snippet and the given dataset that has been stored in the data variable that demonstrates the use of GaussianNB from scikit-learn:
| Outlook | Temperature | Humidity | Windy | Play Golf | |
|---|---|---|---|---|---|
| 0 | Rainy | Hot | High | False | No |
| 1 | Overcast | Hot | High | False | Yes |
| 2 | Sunny | Cool | Normal | False | Yes |
| 3 | Sunny | Cool | Normal | True | No |
| 4 | Overcast | Cool | Normal | True | Yes |
| 5 | Rainy | Cool | Normal | False | Yes |
| 6 | Rainy | Hot | High | True | No |
| 7 | Overcast | Hot | High | False | Yes |
| 8 | Sunny | Cool | Normal | True | No |
| 9 | Overcast | Cool | Normal | True | Yes |
| 10 | Rainy | Cool | Normal | False | Yes |
| 11 | Overcast | Hot | Normal | False | Yes |
from sklearn.preprocessing import OneHotEncoder, LabelEncoderfrom sklearn.naive_bayes import GaussianNB
X = data.drop("Play Golf", axis=1)y = data["Play Golf"]
X = OneHotEncoder(sparse_output=False).fit_transform(X)y = LabelEncoder().fit_transform(y)
estimator = GaussianNB()
estimator.fit(X,y)
print(estimator.class_prior_) # gives the priori of labels(y)What is the prior probability of the label being “No”? i.e.
Correct answer: 0.335 (accepted within ±0.015)
What is the output of the following code?
from sklearn.neighbors import KNeighborsClassifierX = [[2,3], [5,6], [10, 11], [15,16], [20,21]]y = [0, 1, 1, 1, 2]knn = KNeighborsClassifier(n_neighbors=3, metric='euclidean', weights='uniform')knn.fit (X, y)print (knn.predict([[8,9]]))Correct answer: 1
Consider the following code snippet:
from sklearn.datasets import fetch_california_housing, load_irisfrom sklearn.decomposition import PCAfrom sklearn.preprocessing import StandardScaler, PolynomialFeaturesfrom sklearn.pipeline import Pipeline, FeatureUnion
X,y = load_iris(return_X_y= True)
polynomial_transform = PolynomialFeatures(degree=3, interaction_only=False, include_bias=False)
combined_features = FeatureUnion([('poly', polynomial_transform), ('pca', PCA(n_components=2))])
pipeline = Pipeline([('features', combined_features), ('scaler', MinMaxScaler())])
X_transformed = pipeline.fit_transform(X)
print(X_transformed.shape)If the shape of X is , what will be the number of features in X_transformed?
Correct answer: 36
What will be the output of the following code snippet? Assume necessary imports.
X = np.array([[1,6], [-2,0], [-0.25, 3.5]])
pca_transform = PCA(n_components=2)
X_transformed = pca_transform.fit_transform(X)
print(pca_transform.explained_variance_ratio_[0])Correct answer: 1.00
What will be the output of the following code snippet?
from sklearn.linear_model import Perceptron# Sample dataX = [[0, 0], [0, 1], [1, 0], [1, 1]]y = [0, 0, 0, 1]
clf = Perceptron(tol=None, shuffle=False)clf.fit(X, y)
print(clf.predict([[2, 2]]))Correct answer: 1
Consider following code snippet, assume necessary imports:
df=pd.DataFrame(data={"Name":['Akash', 'Brajesh', 'Charu', 'Deepak'], "Maths":[34,43,56,77], "English":[23,45,67,82], "Hindi":[53,35,np.nan,"hi"],}, index = range(11,15))Based on the above data, answer the given subquestions.
Correct answers
Consider following code snippet, assume necessary imports:
df=pd.DataFrame(data={"Name":['Akash', 'Brajesh', 'Charu', 'Deepak'], "Maths":[34,43,56,77], "English":[23,45,67,82], "Hindi":[53,35,np.nan,"hi"],}, index = range(11,15))Based on the above data, answer the given subquestions.
Correct answer: 120
Consider following code snippet, assume necessary imports:
df=pd.DataFrame(data={"Name":['Akash', 'Brajesh', 'Charu', 'Deepak'], "Maths":[34,43,56,77], "English":[23,45,67,82], "Hindi":[53,35,np.nan,"hi"],}, index = range(11,15))Based on the above data, answer the given subquestions.
Correct answer: 3
Consider following code snippet, assume necessary imports:
df=pd.DataFrame(data={"Name":['Akash', 'Brajesh', 'Charu', 'Deepak'], "Maths":[34,43,56,77], "English":[23,45,67,82], "Hindi":[53,35,np.nan,"hi"],}, index = range(11,15))Based on the above data, answer the given subquestions.
What is the output of the following code snippet:
def getPassing(aRow): if aRow['Maths']>=35 and aRow['English']>=35: return True return False
df.apply(getPassing).sum()Enter -1, if you think the given statement will generate an error.
Correct answer: -1
Consider following code snippet, assume necessary imports:
df=pd.DataFrame(data={"Name":['Akash', 'Brajesh', 'Charu', 'Deepak'], "Maths":[34,43,56,77], "English":[23,45,67,82], "Hindi":[53,35,np.nan,"hi"],}, index = range(11,15))Based on the above data, answer the given subquestions.
The number of rows will decrease in the dataset.
The number of columns will decrease in the dataset.
The number of columns and number of rows, both, will decrease in the dataset.
There will be no change.
Insufficient information.
Correct answer
The number of rows will decrease in the dataset.