Table of Contents
Data mining has two complementary goals: understanding and prediction. It encompases many technologies and techniques from AI, physics, statistics, and machine learning.
Data Sets
Data mining operates on observational, secondary, or retrospective data. This is not data gathered for an experiment, but is often recorded for other purposes and is cheap to obtain.
There are common types or sources of data that have similar properties e.g. web data or streaming data.
Properties of Data
- numeric, text, visual
- discrete or continuous
Main Data Mining Techniques
Descriptive Methods
Exploratory Data Analysis Data Visualization Dimensionality reduction Clustering Pattern detection and anomaly detection
Predictive Modelling
The stuff that makes money
Classification Ranking Regression Matrix completion (recommender systems)
Typical Data Mining Process
- Problem formulation
- define goals
- understand the problem domain
- Data Definition
- How does the available data address the problem?
- Validation
- Apply rigorous testing to proposed solutions
- can be considered “productization”
Emerging Fields
Last modified: 2017/05/09 22:19
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 4.0 International