CALL for Papers:
-------------------

Special Issue: 'Evolving Soft Computing Techniques'
http://www.flll.jku.at/sites/default/files/u6/CFP_SI_EvolvingSoftCompTechniques.pdf

Journal Applied Soft Computing
http://www.journals.elsevier.com/applied-soft-computing/


GUEST EDITORS

Abdelhamid Bouchachia
Department of Informatics-Systems
Group of Software Engineering & Soft Computing
University of Klagenfurt, Austria
Email: hamid@isys.uni-klu.ac.at

Edwin Lughofer
Department of Knowledge-Based Mathematical Systems
Johannes Kepler University Linz, Austria
Email: edwin.lughofer@jku.at

Moamar Sayed-Mouchaweh
Ecole des Mines de Douai
Computer Science and Automatic Control Lab, France
Email: moamar.sayed-mouchaweh@mines-douai.fr


IMPORTANT DATES

Submission deadline: September 1st, 2012
First author notification: December 1st, 2012
Revised version: February 1st, 2012
Final notification: April 1st, 2013
Publication: 2013



SCOPE of the ISSUE

In nowadays industrial systems, the necessity of on-line learning becomes more and more an essential aspect, as upcoming new system states, changing operating conditions and environmental influences need to be integrated on demand and on-the-fly into the models. Otherwise, the predictive quality of the models may deteriorate significantly due to severe extrapolation cases. Re-training of the models is usually
not a feasible option whenever data from streams is continuously arriving as not terminating within a reasonable time-frame. Thus, recently a field of research emerged which is addressing this problem by using methodologies with the ability to train and permanently update models in an incremental, step-wise manner. The models equipped with these capabilities are called evolving models.

This special issue aims at laying a bridge between incremental learning methodologies, concepts, techniques and aspects which are basically motivated within the field of machine learning and any type of soft computing model architectures, favoring some sort of interpretability, mimicking human brain modeling and investigating concepts from evolution theory.
This special issue intends to draw a picture of the recent advances and challenges in evolving soft-computing based systems including evolving fuzzy systems, evolving neural networks, dynamic evolutionary algorithms and any evolving hybrid systems (e.g. evolving neuro-fuzzy systems, evolving evolutionary neural networks, dynamic fuzzy evolutionary algorithms, etc.). Particularly, the special issue aims at soliciting contributions dealing with real-world applications that present dynamic facets requiring on-line learning capabilities. The connection of evolving soft computing to specific machine learning and data mining concepts such as active learning,
dynamic feature weighting/selection, drift analysis in data streams, complexity reduction issues, outlier treatment as well as reliability issues are of high relevance.


TOPICS

Evolving fuzzy systems (EFS) including:
o Evolving fuzzy classifiers
o Evolving fuzzy clustering
o Evolving fuzzy regression
o Evolving Takagi-Sugeno-Kang fuzzy systems
o Evolving neuro-fuzzy approaches
o Evolving fuzzy controllers
o Stability, process-safety and computational related aspects
o Complexity reduction and interpretability issues in EFS
o Reliability in model predictions and parameters

Evolving neural networks including:
o Online learning paradigm
o Sequential radial basis functions networks
o Online and incremental support vector machines
o Online perceptron-like neural networks
o Online probabilistic neural networks
o Incremental self-organizing maps
o Stability and plasticity issues
o Issues regarding forgetting

Dynamic Evolutionary Algorithms including:
o Change detection in the environment
o Convergence and computational issues
o Adaptive evolutionary computation
o Methods and strategies of dynamic optimization
o Dynamic multi-objective optimization
o Real-world applications of dynamic optimization

Hybrid methodologies
o incremental genetic fuzzy systems
o evolving neuro-fuzzy approaches
o adaptive neural network training with GAs

Evolving soft computing techniques in connection with
o Active and semi-supervised learning strategies
o Techniques to address “Concept Drift”
o Online/Incremental Feature Selection
o Online tuning via human-machine interaction

Real-World Applications of evolving soft computing techniques
o Online modelling and identification
o Online fault detection and decision support systems
o Online media classification
o Smart systems
o Robotics
o Applications for mining in huge data bases
o Web applications
o Adaptive chemometric models
o Modeling in dynamic processes
o Online time series analysis and stock market forecasting


SUBMISSIONS

Manuscripts should be submitted via the Elsevier Editorial System http://ees.elsevier.com/asoc/. Please choose
“SC: Evolving Techniques” when specifying the Article Type.