CDSP 3 Quiz

1. What is the main purpose of Exploratory Data Analysis (EDA)?

A: To finalize the project deliverables
B: To test machine learning models
C: To maximize insights and reveal underlying patterns in the data
D: To analyze only missing values

2. Which method in pandas can you use to get information about the attributes of a dataset?

A: info()
B: head()
C: describe()
D: tail()

3. What is a target feature in a supervised dataset?

A: The variable that you want to predict or categorize
B: The primary identifier of each row in a dataset
C: A variable used to measure feature relevance
D: A feature that represents missing values

4. What type of bias occurs when a sample doesn’t represent the entire population?

A: Reporting bias
B: Selection bias
C: Measurement bias
D: Observer bias

5. Which technique is used to create new, non-observed data points in a dataset?

A: Random oversampling
B: Random undersampling
C: SMOTE (Synthetic Minority Oversampling Technique)
D: Stratified sampling

6. What is a common cause of imbalanced datasets?

A: Random oversampling
B: Inaccurate labels in the dataset
C: Majority of observations belonging to one class
D: Too many features

7. What is the purpose of a confusion matrix in evaluating a model?

A: To adjust the learning rate
B: To compare true and predicted classes
C: To calculate missing values
D: To analyze feature importance

8. What does the Pearson correlation coefficient measure?

A: The number of missing values
B: The linear correlation between two variables
C: The accuracy of a machine learning model
D: The distribution of categorical data

9. What kind of distribution is indicated by a bell-shaped, symmetrical curve?

A: Normal distribution
B: Skewed distribution
C: Bimodal distribution
D: Multimodal distribution

10. What does a positive value of skewness indicate about a distribution?

A: It is symmetrical
B: The peak is shifted to the left
C: The peak is shifted to the right
D: There is no peak

11. What is one benefit of using box plots in data analysis?

A: They identify the most common value in the dataset
B: They visually depict quartiles and potential outliers
C: They show correlations between two variables
D: They display the mean value of a dataset

12. What does feature scaling aim to achieve in data preprocessing?

A: Create new features from existing ones
B: Standardize feature values to a similar range
C: Remove unnecessary data points
D: Impute missing values

13. Which of the following is a type of transformation function?

A: Imputation
B: One-hot encoding
C: Logarithmic transformation
D: Cross-validation

14. What is the goal of dimensionality reduction?

A: To increase the number of features in a dataset
B: To eliminate redundant or irrelevant features
C: To perform data augmentation
D: To oversample minority classes

15. What is one purpose of feature splitting?

A: To create new datasets from existing ones
B: To separate composite features into multiple, meaningful components
C: To identify missing values in a dataset
D: To impute null values

16. What is a common issue with datasets that have too many features relative to the number of data points?

A: Feature scaling becomes less important
B: The model’s ability to learn patterns may degrade
C: The dataset will contain duplicate rows
D: The data will have more missing values

17. Which of the following is an example of data encoding?

A: Scaling numeric variables
B: Imputing missing values
C: Converting categorical data into numbers
D: Removing duplicates from a dataset

18. What is the advantage of using normalization for feature scaling?

A: It minimizes the effect of outliers
B: It focuses on reducing skewness
C: It only applies to categorical features
D: It increases the dataset size

19. What technique is most useful for handling datasets with a large number of missing values?

A: Feature extraction
B: Random sampling
C: Imputation
D: Dimensionality reduction

20. What is the primary goal of hyperparameter tuning in machine learning models?

A: To adjust the model’s architecture
B: To minimize the training time
C: To optimize the model's performance on the validation data
D: To reduce the amount of data required for training