Fractal Connection Structure: A Simple Way to Improve Generalization in Nonlinear Learning Systems (Special Section on Nonlinear Theory and its Applications)
スポンサーリンク
概要
- 論文の詳細を見る
In this work, introduction of fractal connection structure in non-linear learning systems like multilayer perceptrons as a means of improving its generalization capability in classification problems has been investigated via simulation on sonar data set in underwater target classification problem. It has been found that fractally connected net has better generalization capability compared to the fully connected net and a randomly connected net of same average connectivity for proper choice of fractual dimension which controlls the average connectivity of the net.
- 社団法人電子情報通信学会の論文
- 1996-10-25
著者
-
Sawada Yasuji
Research Center For Electrical Communication Tohoku University
-
CHAKRABORTY Basabi
Research Center for Electrical Communication, Tohoku University
-
Chakraborty Basabi
Research Center For Electrical Communication Tohoku University
関連論文
- Practical Methods of Measuring the Generalized Dimension and the Largest Lyapunov Exponent in High Dimensional Chaotic Systems
- Global Spectral Characterization of Chaotic Dynamics
- Generation Mechanism of a Localized Target Pattern in Electrohydrodynamic Convection
- Optical and Acoustical Investigations of Critical Fluctuation and Critical Slowing for the Convectional Instability
- Fractal Neural Network : Computational Performance as an Associative Memory : General and Mathematical Physics
- Spatial Instabilities and Onset of Chaos in Rayleigh-Benard System with Intermediate Aspect-Ratio
- Study of Hardware Integration of An Artificial Neural Network with A New Analog Memory
- Dynamics and Mechanism of Self-Sustained Oscillation of Lipid Impregnated in a Single Hole
- Diffusion-Limited Aggregation with a Restriction on the Growth Direction
- 4. Growth Pattern in Biological and Nonbiological Systems
- Phase Wave Propagation in the Rectangular Convective Structure of Nematic Liquid Crtstal
- Fractal Neural Network Feature Selector for Automatic Pattern Recognition System (Special Section on Nonlinear Theory and Its Applications)
- Fractal Structure and Dynamics for Kinetic Cluster Aggregation of Polystyrene Colloids
- A New Scheme of Spatio-Temporal Chaos Created by the Interaction between the Phase Waves and the Topological Defects : Complex Dynamics in Nonlinear Systems
- On the Nonlinear Profile of the Heat Pulse in Liquid Helium II.
- Flow at the Onset of Traveling Wave Convection
- Fractal Connection Structure: A Simple Way to Improve Generalization in Nonlinear Learning Systems (Special Section on Nonlinear Theory and its Applications)
- Direct Measurement of the Switching Waveform in a DC-SQUID
- Phase Diagram for Two Weakly Coupled Oscillatory Chemical Systems
- Natural Laws and Information Processing (Special Issue on New Architecture LSIs)
- Analysis of Learning Processes of Chaotic Time Series by Neural Networks
- Rapid patterning in 2-D cultures of Dictyostelium cells and its relationship to zonal differentiation
- Rapid patterning of Dictyostelium discoideum cells under confined geometry and its relation to differentiation