Skip to main content

Unit 2: Subjective Question Bank

1. Define Data Mining.​

Data Mining is process of extracting useful, hidden patterns and knowledge from large datasets. It is the analysis step of KDD.

2. What is KDD? Explain the steps briefly.​

KDD (Knowledge Discovery in Databases) is the full process of finding knowledge in data.

  1. Cleaning: Removing noise/errors.
  2. Integration: Combining multiple data sources.
  3. Selection: Choosing relevant data.
  4. Transformation: Converting data into mining formats (Scaling).
  5. Mining: Extracting patterns (The core step).
  6. Evaluation: Identifying truly useful patterns.
  7. Presentation: Visualizing results for the user.

3. Data Mining vs. Data Warehousing.​

  • Warehouse: A repository used to store and integrate data.
  • Mining: A process used to analyze and find patterns in that data.

4. Major issues in data mining.​

  • Performance: Handling massive data quickly.
  • Data Diversity: Mining complex data like images or web pages.
  • User Interaction: Making mining tools easy to use.
  • Ethics: Protecting user privacy.

5. Data, Information, and Knowledge.​

  • Data: Raw facts (e.g., "50").
  • Information: Data with meaning (e.g., "50 items sold").
  • Knowledge: Information used for action (e.g., "Sales are high; restock now").

6. What is data cleaning?​

The process of removing noise (errors) and handling missing values to improve data quality.

7. What is data transformation?​

Changing data into formats suitable for mining.

  • Normalization: Scaling numbers (e.g., 0 to 1).
  • Aggregation: Summarizing data (e.g., daily to monthly sales).

8. What is a data repository? Give types.​

A centralized place to store data.

  • Relational: Tables with rows/columns.
  • Warehouse: Historical data for analysis.
  • Transactional: Fast, everyday records (OLTP).

9. Classification vs. Clustering.​

  • Classification: Predicting labels for new data (Supervised). E.g., Spam vs. Not Spam.
  • Clustering: Grouping similar items without labels (Unsupervised). E.g., Grouping customers by behavior.

10. What is an association rule?​

Finding items that frequently occur together. Example: People who buy bread also buy butter (Bread →\rightarrow Butter).

11. What are outliers?​

Data points that are significantly different from the rest. Use Case: Used for Fraud Detection.

12. Data mining functionalities with examples.​

  • Association: Finding related items (Market Basket).
  • Classification: Categorizing data (Loan Approval).
  • Clustering: Finding natural groups (News categories).
  • Outlier Analysis: Finding anomalies (Credit card fraud).

13. Descriptive vs. Predictive tasks.​

  • Descriptive: Summarizes current data (e.g., Clustering).
  • Predictive: Forecasts future or unknown values (e.g., Regression).
  • Web Mining: Analyzing website usage.
  • Multimedia Mining: Analyzing images and video.
  • Spatial Mining: Analyzing location/map data.
  • Privacy-Preserving: Mining data without exposing personal details.

15. Methods of presenting models.​

  • IF-THEN Rules: Logical "If this, then that" statements.
  • Decision Trees: A flowchart structure showing tests and outcomes.