Data Mining: Extracting knowledge from data

   

Wednesday 23 February

  
16:30 - 17:25 Lecture 5 Data Mining:  Extracting knowledge from data Petr Olmer

A hidden knowledge can be stored in databases. How to discover it? How can we search for an answer, if we do not know a question? Data mining can help. The objective of the lecture is to introduce basic methods of knowledge discovery in structured data, and also in an unstructured text.

 

1. What and why

  • Data mining, knowledge discovery, data exploration

  • Machine learning

  • Statistics

2. Data mining as a process

  •  CRISP-DM method

  • Predictive and descriptive tasks

  •  Concepts, instances, attributes

3. Models and algorithms

  • Decision trees

  • Classification rules

  • Association rules

  • k-nearest neighbors

  • Cluster analysis

4. Text mining: How does Google News work

  • Converting unstructured text to structured data

  • Cluster analysis