Data Mining — Exam Power Notes
Quick Revision Guide
1. What Is Data Mining?
Data mining discovers useful patterns, relationships, trends, and knowledge from large datasets using ideas from
databases, statistics, and machine learning.
2. KDD Process
The Knowledge Discovery in Databases process commonly includes selection, cleaning, transformation, data mining, and
interpretation or evaluation.
3. Major Data Mining Tasks
Task Purpose
Classification Assigns items to predefined categories.
Clustering Groups similar items without predefined labels.
Association Finds relationships or co-occurring items.
Regression Predicts numerical values.
Anomaly Detection Finds unusual observations.
4. Data Preprocessing
Raw data may contain missing values, duplicates, noise, and inconsistent formats. Cleaning and transformation improve
data quality.
5. Classification
A supervised task using known labels. Common methods include decision trees, k-nearest neighbors, and naive Bayes.
6. Clustering
An unsupervised technique that groups observations by similarity. K-means is a common example.
7. Association Rules
Association analysis discovers item relationships. Support, confidence, and lift are common measures.
8. Decision Trees
A tree represents tests and outcomes. Internal nodes are tests, branches are outcomes, and leaves contain predictions.
9. Overfitting
A model overfits when it learns training data too closely and performs poorly on unseen data. Validation and regularization
can help.
10. Evaluation
Classification metrics include accuracy, precision, recall, and F1-score. Metric choice depends on the task.
11. Applications
Data mining supports customer analysis, fraud detection, recommendations, scientific research, business intelligence, and
many other fields.
12. Ethical Issues
Quick Revision Guide
1. What Is Data Mining?
Data mining discovers useful patterns, relationships, trends, and knowledge from large datasets using ideas from
databases, statistics, and machine learning.
2. KDD Process
The Knowledge Discovery in Databases process commonly includes selection, cleaning, transformation, data mining, and
interpretation or evaluation.
3. Major Data Mining Tasks
Task Purpose
Classification Assigns items to predefined categories.
Clustering Groups similar items without predefined labels.
Association Finds relationships or co-occurring items.
Regression Predicts numerical values.
Anomaly Detection Finds unusual observations.
4. Data Preprocessing
Raw data may contain missing values, duplicates, noise, and inconsistent formats. Cleaning and transformation improve
data quality.
5. Classification
A supervised task using known labels. Common methods include decision trees, k-nearest neighbors, and naive Bayes.
6. Clustering
An unsupervised technique that groups observations by similarity. K-means is a common example.
7. Association Rules
Association analysis discovers item relationships. Support, confidence, and lift are common measures.
8. Decision Trees
A tree represents tests and outcomes. Internal nodes are tests, branches are outcomes, and leaves contain predictions.
9. Overfitting
A model overfits when it learns training data too closely and performs poorly on unseen data. Validation and regularization
can help.
10. Evaluation
Classification metrics include accuracy, precision, recall, and F1-score. Metric choice depends on the task.
11. Applications
Data mining supports customer analysis, fraud detection, recommendations, scientific research, business intelligence, and
many other fields.
12. Ethical Issues
