Data Science and Big Data Analytics

Type: 
Lecture

Dates, Modules, etc.

  • Lecture: Tuesdays, 14:15-15:45 o'clock, Provisorischer Hörsaal B (Chemie)
  • Exercise: Thursdays, 13:15-14:45 o'clock, Informatik-Provisorium - 0.103 - Date of first session will be announced via StudIP. Details will be given in the first lecture. We may extend the time slot based on the number of participants.
  • Module: M.Inf.1151, 2.17 Data Warehousing and Data Mining Techniques (ITIS)
  • The lecture will NOT be available via Webstream!

Registration

This lecture requires registration. The registration procedure will be explained during the first lecture. Registration and active participation in a group project is mandatory in order to be allowed to participate in the final exam.

Description

The main topic of this lecture is data science, i.e., methods to extract information from data with a scientific approach. We approach this topic from a practical side in this lecture. This means, that we concern ourselves directly with what algorithms do, and where they should be applied. The details of the algorithms and the theory behind them are not part of this lecture. Methods considered in this lecture include:

  • Association rule mining with the APRIORI approach
  • Clustering with k-means, EM for gaussian mixtures, DBSCAN, and single linkage clustering
  • Classification with k-nearest neighbor, decision trees, random forests, logistic regression, naive Bayes, support vector machines, and neural networks
  • Linear regression with ridge and lasso
  • Time series analysis with ARMA
  • Fundamentals of text mining

Additionally, we will consider the analysis of Big Data. In this context, we will consider the following topics:

  • The MapReduce paradigm
  • Apache Hadoop and Apache Spark

Materials

To be announced.

2011 © Software Engineering For Distributed Systems Group

Main menu 2