Comparative Study of Dimension Reduction Approaches With Respect to Visualization in 3-Dimensional Space

<p>In the present big data era, there is a need to process large amounts of unlabeled data and find some patterns in the data to use it further. If data has many dimensions, it is very hard to get any insight of it. It is possible to convert high-dimensional data to low-dimensional data using different techniques, this dimension reduction is important and makes tasks such as classification, visualization, communication and storage much easier. The loss of information should be less while mapping data from high-dimensional space to low-dimensional space. Dimension reduction has been a significant problem in many fields as it needs to discard features that are unimportant and discover only the representations that are needed, hence it gathers our interest in this problem and basis of the research. We consider different techniques prevailing for dimension reduction like PCA (Principal Component Analysis), SVD (Singular Value Decomposition), DBN (Deep Belief Networks) and Stacked Auto-encoders. This thesis is intended to ultimately show which technique performs best for dimension reduction with the help of studied experiments.</p>

Name thesis:degree_name

Master of Science in Computer Science (MSCS)

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Year dc:date.available

2016

Author dc:creator

• Chenna, Pooja

Contributors dc:contributor

• Ying Xie
• Frank Tsui
• Selena He

• dimension reduction
• PCA
• SVD
• deep learning
• stacked auto-encoders
• DBN
• visualization
• entropy
• Other Computer Engineering

Repository record dc:identifier

https://digitalcommons.kennesaw.edu/cs_etd/3

OAI identifier oai:identifier

oai:digitalcommons.kennesaw.edu:cs_etd-1003