Reducing the dimensionality of data using different techniques

Main Article Content

Ahmed Fadhil Mohammed, Dhulfiqar Abbas Hoiji

Abstract

High-dimensional data processing is really the key issue in many systems, including contentbased
extraction, voice signals, fMRI analyses, electroencephalogram object detection,
multimedia extraction, market-based technologies, etc. In order to enhance the
system's effectiveness, the data dimensions must be minimized to a low dimensional space.
Inthis paper, we analyzed linearization, nonlinear and network embedding dimensionality
reduction methods. A few of these methods are ideal and used for linear data that have linear
relationship between data points, and many other dimensionality reduction methods are used for
nonlinear data that have nonlinear relationship among data points.From an analysis of this paper,
we found that Structural Deep Network Embedding (SDNE), LINE (Large-Scale Network
Embedding) and Nod2Vec are the best techniques for dimensionality reduct ion in network
data.Furthermore, every approach has its own characteristics and drawbacks. This study
presents different methods utilized to minimize the high dimensional data into low dimensional
space.

Downloads

Download data is not yet available.

Metrics

Metrics Loading ...

Article Details

How to Cite
Ahmed Fadhil Mohammed, Dhulfiqar Abbas Hoiji. (2021). Reducing the dimensionality of data using different techniques. Turkish Journal of Computer and Mathematics Education (TURCOMAT), 11(1), 366–373. https://doi.org/10.17762/turcomat.v11i1.11385
Section
Research Articles