Data Mining · CSC213

What is Data Mining? Tasks: Classification, Clustering, Regression, Association Rules

Data Mining Study guides

Data mining is the extraction of interesting, non-obvious and useful patterns or knowledge from huge amounts of data. Its tasks are split into predictive tasks and descriptive tasks.

Data mining became a necessity because we collect far more data than people can inspect by hand: transactions, sensors, web logs, medical records. Mining finds patterns that support decisions, such as which customers are likely to leave or which transactions look fraudulent.

Predictive tasks use some variables to predict unknown or future values of another variable. Classification predicts a category (will this customer churn: yes or no), and regression predicts a continuous number (tomorrow's sales). Descriptive tasks find human-interpretable patterns that describe the data. Clustering groups similar objects without labels and is a form of unsupervised learning. Association rule discovery finds items that occur together (customers who buy bread often buy butter). Anomaly or outlier detection finds objects that do not follow the general behaviour of the data, such as credit card fraud.

A quick test: if the task looks up stored values, sorts a database or computes a simple total, it is ordinary query processing, not data mining. If it discovers a pattern or predicts something not directly stored, it is a data mining task.

Key points

  • Two families: predictive (classification, regression) and descriptive (clustering, association rules, anomaly detection).
  • Classification predicts a label; regression predicts a continuous value.
  • Clustering is unsupervised: no labels are given.
  • Association rules find co-occurring items.
  • Anomaly detection looks for objects that differ from the general model, like fraud.

Worked example

Dividing customers of a company by their purchase behaviour is clustering. Predicting tomorrow's temperature from past data is regression. Finding that diapers and beer are bought together is association rule discovery. Sorting a student table by ID is just a database operation.

Common questions

Is clustering supervised or unsupervised?

Unsupervised, because the groups are discovered without predefined labels.

What is the difference between classification and regression?

Classification predicts a category. Regression predicts a continuous numeric value.

Practice questions on this topic

Try these, then open the app to check your answers with explanations.

  1. The following activity is a data mining Task: Predicting the future stock price of a company using historical records.
    • A. false
    • B. true
  2. Functions of Data Mining are
    • A. Prediction and characterization
    • B. Association and correctional analysis classification
    • C. Cluster analysis and evolution analysis
    • D. All of the above
  3. The following activity is a data mining Task: Monitoring the heart rate of a patient for abnormalities.
    • A. false
    • B. true
  4. The following activity is a data mining Task: Predicting the outcomes of tossing a (fair) pair of dice.
    • A. false
    • B. true
  5. The following activity is a data mining Task: Dividing the customers of a company according to their gender.
    • A. false
    • B. true
  6. The following activity is a data mining Task: Sorting a student database based on student identification numbers.
    • A. false
    • B. true