wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

MSTIP - FinalExam - Data Mining

Total questions: 100

Worksheet time: 2hrs 40mins

Name
Class
Date
1.

Which of the following is NOT a typical application of data mining?

a)

Fraud detection

b)

Weather forecasting

c)

Market basket analysis

d)

Customer segmentation

2.

What is the main difference between data mining and traditional data analysis?

a)

Data mining is faster

b)

Data mining uses statistical methods only

c)

Data mining discovers hidden patterns automatically

d)

Data mining replaces databases

3.

Which of the following is a descriptive data mining task?

a)

Classification

b)

Regression

c)

Clustering

d)

Prediction

4.

Which term refers to the process of cleaning and transforming raw data before mining?

a)

Data mining

b)

Data integration

c)

Data pre-processing

d)

Data warehousing

5.

In classification, the output variable is:

a)

Continuous

b)

Categorical

c)

Text

d)

Unsupervised

6.

Which of the following is an example of a supervised learning task in data mining?

a)

Clustering

b)

Association rule mining

c)

Classification

d)

Dimensionality reduction

7.

What is the purpose of association rule mining?

a)

Predicting a numeric value

b)

Discovering relationships among variables in large databases

c)

Segmenting similar customers

d)

Reducing data dimensions

8.

The term “support” in association rule mining refers to:

a)

The number of items in a transaction

b)

The number of rules generated

c)

The frequency of occurrence of an itemset

d)

The accuracy of classification

9.

Which technique is best suited for detecting unusual patterns that do not conform to expected behavior?

a)

Clustering

b)

Classification

c)

Outlier detection

d)

Association rules

10.

Which data mining technique involves grouping similar data points without predefined labels?

a)

Classification

b)

Clustering

c)

Regression

d)

Decision tree

11.

A confusion matrix is used to evaluate the performance of:

a)

Clustering

b)

Classification models

c)

Association rule mining

d)

Regression analysis

12.

What does “overfitting” refer to in data mining?

a)

The model fits training data well but performs poorly on new data

b)

The model does not fit the training data well

c)

The model predicts only categorical values

d)

The data is too large to process

13.

Which of the following is a characteristic of a good data mining algorithm?

a)

It should work only with numeric data

b)

It should overfit the data

c)

It should generalize well to new data

d)

It should ignore outliers

14.

Which of these is a common algorithm for classification?

a)

Apriori

b)

K-Means

c)

Decision Trees

d)

PCA

15.

The K in the K-Means algorithm represents:

a)

Number of rules

b)

Number of neighbors

c)

Number of clusters

d)

Number of attributes

16.

Which data mining task is best for predicting a numerical outcome?

a)

Classification

b)

Regression

c)

Clustering

d)

Association

17.

In data warehousing, the process of collecting and managing data from varied sources is called:

a)

Data mining

b)

Data integration

c)

Data pruning

d)

Data visualization

18.

Principal Component Analysis (PCA) is primarily used for:

a)

Classification

b)

Regression

c)

Dimensionality reduction

d)

Clustering

19.

Which one of the following is NOT a typical step in the data mining process?

a)

Data collection

b)

Data preprocessing

c)

Modeling

d)

Software debugging

20.

What does the term “curse of dimensionality” refer to?

a)

Too little data to analyze

b)

Data with too many missing values

c)

Increased data volume reducing model performance

d)

Increased number of attributes causing performance issues

21.

Which of the following is an example of a classification problem in supervised learning?

a)

Predicting house prices

b)

Forecasting stock values

c)

Identifying spam emails

d)

Estimating rainfall amounts

22.

In supervised learning, the output variable is:

a)

Unlabeled

b)

Always categorical

c)

Labeled

d)

Always continuous

23.

Which metric is most appropriate for evaluating a regression model?

a)

Accuracy

b)

F1 Score

c)

Mean Absolute Error (MAE)

d)

Precision

24.

Which algorithm is primarily used for classification tasks?

a)

Linear Regression

b)

Logistic Regression

c)

K-Means Clustering

d)

Principal Component Analysis

25.

Overfitting in a model refers to:

a)

High training error, low test error

b)

Low training error, high test error

c)

Equal error in training and test sets

d)

Model ignoring noise in training data

26.

What is the purpose of cross-validation?

a)

To reduce dimensionality

b)

To improve training speed

c)

To evaluate model performance on unseen data

d)

To normalize data

27.

Which of the following is not a supervised learning algorithm?

a)

Decision Tree

b)

Naive Bayes

c)

Support Vector Machine

d)

Apriori

28.

In a confusion matrix, a false positive means:

a)

The model predicted negative, but it was positive

b)

The model predicted positive, but it was negative

c)

Both prediction and actual were negative

d)

Both prediction and actual were positive

29.

What does the "bias-variance trade-off" describe?

a)

Trade-off between training and test data

b)

Trade-off between number of features and samples

c)

Trade-off between underfitting and overfitting

d)

Trade-off between classification and regression

30.

Which supervised learning algorithm is most sensitive to feature scaling?

a)

Decision Trees

b)

K-Nearest Neighbors

c)

Naive Bayes

d)

Random Forest

31.

Which of the following is a disadvantage of k-NN?

a)

It is parametric

b)

Training is very slow

c)

It doesn't support multi-class classification

d)

It cannot be used for regression

32.

The ROC curve is useful for evaluating:

a)

Regression models

b)

Clustering algorithms

c)

Binary classifiers

d)

Dimensionality reduction

33.

Which of the following helps prevent overfitting in decision trees?

a)

Increasing depth

b)

Increasing features

c)

Pruning

d)

Removing cross-validation

34.

Which method combines predictions from multiple models to improve accuracy?

a)

Feature selection

b)

Ensemble learning

c)

Regularization

d)

Cross-validation

35.

Which algorithm is a type of ensemble method?

a)

Naive Bayes

b)

KNN

c)

Random Forest

d)

Logistic Regression

36.

The learning rate in gradient descent:

a)

Controls how much we adjust weights with respect to the loss

b)

Determines the model's accuracy

c)

Specifies the number of features

d)

Is fixed in all models

37.

L1 regularization encourages:

a)

Grouping of correlated features

b)

Smooth decision boundaries

c)

Sparse feature selection

d)

Equal feature importance

38.

Which of the following can help improve the generalization of a model?

a)

Using the training set for evaluation

b)

Increasing model complexity

c)

Data augmentation

d)

Removing regularization

39.

Which loss function is commonly used in logistic regression?

a)

Mean Squared Error

b)

Hinge Loss

c)

Cross-Entropy Loss

d)

L1 Loss

40.

What is the main objective of supervised learning?

a)

Group similar items

b)

Explore data without labels

c)

Predict output values from labeled inputs

d)

Reduce dimensionality

41.

Which of the following is a typical use case of unsupervised learning?

a)

Predicting stock prices

b)

Classifying spam emails

c)

Customer segmentation

d)

House price prediction

42.

In clustering, what is the primary goal?

a)

To predict future outcomes

b)

To group similar instances together

c)

To reduce dimensionality

d)

To classify data using labeled examples

43.

Which of the following algorithms is commonly used for clustering?

a)

K-Means

b)

Linear Regression

c)

Naive Bayes

d)

Decision Tree

44.

In Principal Component Analysis (PCA), what is the principal component?

a)

The first variable in the dataset

b)

A linear combination of original features

c)

A decision boundary

d)

A cluster center

45.

What metric does the K-Means algorithm minimize?

a)

Mean Absolute Error

b)

Euclidean distance from centroids

c)

Entropy

d)

Gini index

46.

Which of the following is a method for dimensionality reduction?

a)

K-Nearest Neighbors

b)

PCA

c)

Random Forest

d)

Logistic Regression

47.

The elbow method is used to determine:

a)

Feature importance

b)

Optimal number of clusters in K-Means

c)

Outliers

d)

Classification accuracy

48.

Hierarchical clustering differs from K-Means in that:

a)

It requires predefined number of clusters

b)

It only works with supervised data

c)

It creates a dendrogram

d)

It is faster for large datasets

49.

Which of the following statements is true about DBSCAN?

a)

It assumes spherical clusters

b)

It requires specifying the number of clusters

c)

It is sensitive to noise

d)

It can find clusters of arbitrary shapes

50.

What is the role of the 'epsilon' parameter in DBSCAN?

a)

Defines the number of clusters

b)

Specifies the minimum variance

c)

Defines the neighborhood radius

d)

Determines the step size

51.

Which of the following is not typically an unsupervised learning technique?

a)

K-Means

b)

PCA

c)

DBSCAN

d)

Linear Discriminant Analysis (LDA)

52.

PCA transforms the data to:

a)

A lower-dimensional space with dependent features

b)

A higher-dimensional space

c)

A lower-dimensional space with uncorrelated features

d)

The original feature space with normalized values

53.

A key disadvantage of K-Means is:

a)

It can find non-linear boundaries

b)

It is deterministic

c)

It requires labeled data

d)

It assumes clusters are spherical and equal in size

54.

Which evaluation metric is used for clustering when labels are unavailable?

a)

Accuracy

b)

Adjusted Rand Index

c)

Silhouette Score

d)

ROC-AUC

55.

Which algorithm is best suited for discovering clusters with varying density?

a)

K-Means

b)

PCA

c)

DBSCAN

d)

Hierarchical clustering

56.

Which of the following is true for hierarchical clustering?

a)

It cannot be visualized

b)

It doesn’t require a distance metric

c)

It always produces the same number of clusters

d)

It can be agglomerative or divisive

57.

The curse of dimensionality affects:

a)

Linear regression performance

b)

K-Means efficiency and distance accuracy

c)

Label encoding in supervised models

d)

Ensemble model performance

58.

What does t-SNE do?

a)

Clusters data points

b)

Classifies data

c)

Projects high-dimensional data into 2 or 3 dimensions for visualization

d)

Removes noise from data

59.

In PCA, what determines how many components should be selected?

a)

Number of classes

b)

Explained variance ratio

c)

Number of original features

d)

Training error

60.

Which technique can detect anomalies in data using unsupervised learning?

a)

K-Nearest Neighbors

b)

DBSCAN

c)

Logistic Regression

d)

Random Forest

61.

In association rule mining, what does the "support" of a rule indicate?

a)

The strength of the rule

b)

The proportion of transactions that contain the itemset

c)

The frequency of the consequent

d)

The confidence level of the rule

62.

Which of the following measures is used to evaluate the reliability of a rule?

a)

Support

b)

Confidence

c)

Lift

d)

All of the above

63.

If an itemset has high support but low confidence, it means:

a)

The rule is strong

b)

The rule is rare

c)

The rule occurs frequently but not reliably

d)

The rule is invalid

64.

What does "lift" measure in association rules?

a)

Probability of consequent given antecedent

b)

Ratio of observed support to expected support under independence

c)

Support of the antecedent

d)

Confidence of the rule

65.

In the Apriori algorithm, what property helps reduce the number of candidate itemsets?

a)

Downward closure property

b)

Monotonicity

c)

Lift maximization

d)

Confidence boosting

66.

Which algorithm improves upon Apriori by reducing the number of database scans?

a)

FP-Growth

b)

Eclat

c)

A-PrioriTid

d)

Naïve Bayes

67.

What is a key difference between Apriori and FP-Growth?

a)

Apriori uses a tree structure

b)

FP-Growth uses candidate generation

c)

Apriori is faster on sparse data

d)

FP-Growth avoids candidate generation

68.

Which of the following is not an association rule evaluation metric?

a)

Conviction

b)

Cosine

c)

Entropy

d)

Lift

69.

What is the "confidence" of the rule A → B if support(A ∪ B) = 0.2 and support(A) = 0.4?

a)

0.1

b)

0.5

c)

0.8

d)

0.6

70.

The FP-tree structure is used to:

a)

Compress the transaction database

b)

Replace itemsets with item graphs

c)

Increase the number of candidate itemsets

d)

Sort rules in descending lift

71.

In association rule mining, what is a “frequent itemset”?

a)

An itemset that occurs less than 50% of the time

b)

An itemset with a support above a user-defined threshold

c)

A set with high confidence

d)

A rule with low lift

72.

Which of the following best describes the “downward closure” property?

a)

All subsets of a frequent itemset must also be frequent

b)

All supersets of an infrequent itemset must be infrequent

c)

Both A and B

d)

None of the above

73.

In the rule A → B, which itemset is the antecedent?

a)

A

b)

B

c)

A ∪ B

d)

B ∪ C

74.

The rule A → B is more interesting if:

a)

Lift ≈ 1

b)

Confidence ≈ 0

c)

Support is very low

d)

Lift > 1

75.

Which approach does FP-Growth use to mine frequent patterns?

a)

Bottom-up search

b)

Top-down search

c)

Divide-and-conquer

d)

Breadth-first search

76.

What is the main disadvantage of the Apriori algorithm?

a)

Requires sorting

b)

Generates too many candidate itemsets

c)

Cannot find frequent patterns

d)

Works only for binary data

77.

Conviction is used in association rule mining to measure:

a)

Support strength

b)

Rule reliability if consequent is false

c)

Co-occurrence

d)

Lift ratio

78.

Which of the following is an example of an association rule?

a)

{milk, bread}

b)

milk → bread

c)

Support({milk, bread}) = 0.4

d)

{milk, butter, eggs}

79.

What type of data is typically used in market basket analysis?

a)

Sequential

b)

Time-series

c)

Transactional

d)

Graph-based

80.

Which of the following would indicate an interesting rule?

a)

Low support, low confidence, lift = 1

b)

High support, low lift

c)

High confidence, lift > 1

d)

Confidence = support

81.

Which metric is best used when the classes are imbalanced?

a)

Accuracy

b)

Precision

c)

Recall

d)

F1 Score

82.

What does a high precision score indicate?

a)

Few false positives

b)

Few false negatives

c)

High true positives

d)

Low accuracy

83.

Which metric is calculated as: TP / (TP + FN)?

a)

Precision

b)

Recall

c)

Specificity

d)

Accuracy

84.

What is the formula for accuracy?

a)

TP / (TP + FN)

b)

TP / (TP + FP)

c)

(TP + TN) / (TP + TN + FP + FN)

d)

TN / (TN + FP)

85.

Which of the following is not an evaluation metric?

a)

Confusion Matrix

b)

Precision

c)

Gradient

d)

ROC AUC

86.

What does ROC curve plot?

a)

Precision vs Recall

b)

True Positive Rate vs False Positive Rate

c)

Accuracy vs Loss

d)

Recall vs Specificity

87.

Which metric is used to evaluate the ranking ability of a classifier?

a)

F1 Score

b)

ROC AUC

c)

Accuracy

d)

Precision

88.

What does a confusion matrix help you visualize?

a)

Model complexity

b)

Prediction errors

c)

Training time

d)

Number of epochs

89.

If a model has high recall but low precision, what does it mean?

a)

Many relevant items are missed

b)

Many irrelevant items are classified as relevant

c)

High number of true negatives

d)

Perfect accuracy

90.

Which metric combines both precision and recall?

a)

Accuracy

b)

Specificity

c)

F1 Score

d)

ROC AUC

91.

Specificity is defined as:

a)

TN / (TN + FP)

b)

TP / (TP + FN)

c)

TP / (TP + FP)

d)

FP / (FP + TN)

92.

Which metric is best for minimizing false negatives?

a)

Precision

b)

Specificity

c)

Recall

d)

F1 Score

93.

In a confusion matrix, which value represents false positives?

a)

Top-left

b)

Top-right

c)

Bottom-left

d)

Bottom-right

94.

What does an AUC score of 0.5 indicate?

a)

Excellent model

b)

Random guessing

c)

Perfect recall

d)

High accuracy

95.

Which metric increases if you increase the threshold in a classifier?

a)

Recall

b)

False Positive Rate

c)

Precision

d)

F1 Score

96.

What happens to recall when the decision threshold is lowered?

a)

Increases

b)

Decreases

c)

Remains same

d)

Cannot be determined

97.

Which scenario best suits using the Precision metric?

a)

Detecting cancer cases

b)

Spam email detection

c)

Face recognition with many classes

d)

Weather forecasting

98.

What is the primary weakness of accuracy in imbalanced datasets?

a)

Too complex to calculate

b)

Ignores true positives

c)

Doesn’t distinguish between classes

d)

Misleading when one class dominates

99.

What is the ideal value of the F1 Score?

a)

0

b)

0.5

c)

1

d)

Depends on model

100.

Which metric would you use if false positives are costlier than false negatives?

a)

Recall

b)

Accuracy

c)

Precision

d)

F1 Score