Dimensionality Reduction in Data Mining Using Artificial Neural Networks

Methodology ◽  
2009 ◽  
Vol 5 (1) ◽  
pp. 26-34
Author(s):  
Rafael Jiménez ◽  
Elena Gervilla ◽  
Albert Sesé ◽  
Juan José Montaño ◽  
Berta Cajal ◽  
...  
Keyword(s):  
Data Mining ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Dimensionality Reduction ◽  
Principal Component ◽  
Personality Variables ◽  
Reduction Techniques ◽  
Graphic Analysis ◽  
Artificial Neural ◽  
Visual Graphic

The use of classic dimension reduction techniques can be considered customary practice within the context of data mining (DM). Nevertheless, although artificial neural networks (ANNs) are one of the most important DM techniques, specific ANN architectures for dimensionality reduction, such as the principal components analysis ANN (PCA-ANN) and the linear auto-associative ANN (LA-ANN), are used on far fewer occasions. In this study, categorical principal component analysis (CATPCA) and the two ANN procedures are studied and compared searching for uniqueness in an applied context relative to personality variables and drug consumption. A sample of 7,030 adolescents completed a personality test made up of 20 dichotomous items with a hypothesized four-factor latent model. Results point out that both ANN factor solutions converge to those obtained using CATPCA. Nevertheless, possible drawbacks of the ANN techniques lie in their relatively complex application, as well as in the need to use visual graphic analysis as a support for interpreting the factorized solutions.

Hybrid Wavelet-Neuro-Fuzzy Systems of Computational Intelligence in Data Mining Tasks

2017 ◽  
pp. 787-825
Author(s):  
Yevgeniy Bodyanskiy ◽  
Olena Vynokurova ◽  
Oleksii Tyshchenko
Keyword(s):  
Data Mining ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Learning Process ◽  
Computational Intelligence ◽  
Group Method ◽  
Learning Criterion ◽  
Adjustment Procedure ◽  
Artificial Neural ◽  
Input Variables

This work is devoted to synthesis of adaptive hybrid systems based on the Computational Intelligence (CI) methods (especially artificial neural networks (ANNs)) and the Group Method of Data Handling (GMDH) ideas to get new qualitative results in Data Mining, Intelligent Control and other scientific areas. The GMDH-artificial neural networks (GMDH-ANNs) are currently well-known. Their nodes are two-input N-Adalines. On the other hand, these ANNs can require a considerable number of hidden layers for a necessary approximation quality. Introduced Q-neurons can provide a higher quality using the quadratic approximation. Their main advantage is a high learning rate. Universal approximating properties of the GMDH-ANNs can be achieved with the help of compartmental R-neurons representing a two-input RBFN with the grid partitioning of the input variables' space. An adjustment procedure of synaptic weights as well as both centers and receptive fields is provided. At the same time, Epanechnikov kernels (their derivatives are linear to adjusted parameters) can be used instead of conventional Gauss functions in order to increase a learning process rate. More complex tasks deal with stochastic time series processing. This kind of tasks can be solved with the help of the introduced adaptive W-neurons (wavelets). Learning algorithms are characterized by both tracking and smoothing properties based on the quadratic learning criterion. Robust algorithms which eliminate an influence of abnormal outliers on the learning process are introduced too. Theoretical results are illustrated by multiple experiments that confirm the proposed approach's effectiveness.

scholarly journals Artificial Neural Networks for Automated Cell Quantification in Lensless LED Imaging Systems

Proceedings ◽  
2018 ◽  
Vol 2 (13) ◽  
pp. 989 ◽  
Author(s):  
Agus Budi Dharmawan ◽  
Gregor Scholz ◽  
Shinta Mariana ◽  
Philipp Hörmann ◽  
Igi Ardiyanto ◽  
...  
Keyword(s):  
Neural Networks ◽  
Principal Component Analysis ◽  
Artificial Neural Networks ◽  
Light Emitting Diode ◽  
Principal Component ◽  
Component Analysis ◽  
Detection Accuracy ◽  
Cell Detection ◽  
Artificial Neural ◽  
Feature Values

Cell registration by artificial neural networks (ANNs) in combination with principal component analysis (PCA) has been demonstrated for cell images acquired by light emitting diode (LED)-based compact holographic microscopy. In this approach, principal component analysis was used to find the feature values from cells and background, which would be subsequently employed as neural inputs into the artificial neural networks. Image datasets were acquired from multiple cell cultures using a lensless microscope, where the reference data was generated by a manually analyzed recording. To evaluate the developed automatic cell counter, the trained system was assessed on different data sets to detect immortalized mouse astrocytes, exhibiting a detection accuracy of ~81% compared with manual analysis. The results show that the feature values from principal component analysis and feature learning by artificial neural networks are able to provide an automatic approach on the cell detection and registration in lensless holographic imaging.

Sign in / Sign up

Export Citation Format

Share Document