Skip to content. | Skip to navigation

Personal tools
You are here: Home Theses Visual Analytics Framework for High Dimensional Data Streams


Prof. Dr. S. Decker
RWTH Aachen
Informatik 5
Ahornstr. 55
D-52056 Aachen
Tel +49/241/8021501
Fax +49/241/8022321

How to find us

Annual Reports





Visual Analytics Framework for High Dimensional Data Streams

Thesis type
  • Master
Student Khan Hafizur Rahman
Status Finished
Submitted in 2018

The aim of this thesis is to develop an effective visualization framework for Stream data. In this framework, we can select the high dimensional datasets from different domains and reduce the dimensions through dimensionality reduction algorithm. For dimensionality reduction, we will use the incremental version of traditional dimensionality reduction algorithm known as Linear Discriminant Analysis. Another prime concern will be regarding the information loss due to the dimensionality reduction. Our framework will be evaluated regarding the information loss and also some more evaluation criteria. The visualisation will be presented to the end user through HeatMap; one of the most used prominent tool for high dimensional data visualization technique

There are two different ways to reduce dimensions from the original dataset. The reduction can be done either by selecting significant features and form a subset of the original set. The other way to reduce it by transforming dimension in order to get a new one or reduced set of dimensions. The first one is called the Feature Selection which is greedy in nature where the later one is known as Feature Extraction. Due to the nature of Feature Selection, it is always a challenge an optimal solution or the subset based on some dened criteria. Orthogonal Centroid Algorithm is a Feature Selection Algorithm. 


Document Actions