Florian Neukart

Publications

• Cortical Artificial Neural Networks and their Evolution - Consciousness-inspired Data Mining

  Florian Neukart, Sorin-Aurel Moraru, Costin-Marius Grigorescu, Peter Szakacs-Simon

  OPTIM 2012 - 13th International Conference on Optimization of Electrical and Electronic Equipment; 05/2012

  When trying to solve classification or time-series prediction problem statements by the application of Artificial Neural Networks (ANNs), commonly applied structures like feed forward or recurrent Multi-Layer Perceptrons (MLP) characteristically tend to come up with bad performance and accuracy. Thi... [more] When trying to solve classification or time-series prediction problem statements by the application of Artificial Neural Networks (ANNs), commonly applied structures like feed forward or recurrent Multi-Layer Perceptrons (MLP) characteristically tend to come up with bad performance and accuracy. This is especially the case when dealing with manifold datasets containing numerous input (predictors) and/or target-attributes and independent from the applied learning methods, activation functions, biases, etc... The cortical ANN, inspired by theoretical aspects of the human consciousness and its signal processing, is an ANN structure having been developed during the research phase of the SHOCID project. Due to its structure, redundancy and error-tolerance is being created, which helps to elude the latterly mentioned problems. Within this elaboration, the cortical ANN is being introduced, as well as an algorithm for evolving this special ANN types' structure until the most suitable solution has been detected.

• Accuracy through complexity – one step further in time-series prediction and classification

  Florian Neukart

  05/2012;

  When trying to solve classification or time-series prediction problem statements by the application of Artificial Neural Networks (ANNs), commonly applied structures like feed forward or recurrent Multi-Layer Perceptrons (MLP) characteristically tend to come up with bad performance and accuracy. The... [more] When trying to solve classification or time-series prediction problem statements by the application of Artificial Neural Networks (ANNs), commonly applied structures like feed forward or recurrent Multi-Layer Perceptrons (MLP) characteristically tend to come up with bad performance and accuracy. The introduced recurrent ANN is a development based on the well-known recurrent ANNs from Jordan and Elman. As the name indicates, it has been developed during the research phase of the SHOCID project, which it owes its name (SHOCID RANN or SRANN). Due to its structure, the memory of commonly applied recurrent ANN structures is being exceeded, so that time series prediction is more effective especially when presenting manifold predictors. Within this elaboration, the SRANN is being introduced, as well as an algorithm for evolving this special ANN types' structure until the most suitable solution has been detected

• Transgenetic NeuroEvolution

  Florian Neukart, Sorin-Aurel Moraru, Costin-Marius Grigorescu, Peter Szakacs-Simon

  OPTIM 2012 - 13th International Conference on Optimization of Electrical and Electronic Equipment; 05/2012

  Transgenetic algorithms can be used for performing a stochastic search by simulating endosymbiotic interactions between a host and a population of endosymbionts as well as information exchange between the host and endosymbionts by agents. The already introduced, computationally intelligent Data Mini... [more] Transgenetic algorithms can be used for performing a stochastic search by simulating endosymbiotic interactions between a host and a population of endosymbionts as well as information exchange between the host and endosymbionts by agents. The already introduced, computationally intelligent Data Mining system "System applying High Order Computational Intelligence in Data Mining” (SHOCID) applies such for Artificial Neural Network (ANN) learning by the combination of one of its learning approaches with a host organism, serving as genetic pool, and transgenetic vectors. The application of an algorithm combining horizontal gene transfer between a host and a symbiont is a completely new ANN learning approach, which increases both learning performance and accuracy to a considerable degree. A further advantage is that the application of transgenetic vectors massively increases the chance of reaching the desired stopping criteria (like a minimum Root Mean Squared Error [RMSE]) instead of abort criteria (like the evolutionary stop after 5,000 generations without improvement although the desired have not been fulfilled), as even learning algorithms like back propagation cannot oscillate or get stuck in local minima due to the inescapable transfer of host genetic material.

• High Order Computational Intelligence in Data Mining - A generic approach to systemic intelligent Data Mining

  Florian Neukart, Costin-Marius Grigorescu, Sorin-Aurel Moraru

  6th International Conference on Speech Technology and Human-Computer Dialogue; 01/2011

  Within this elaboration a generic system, subsequently referred to as System applying High Order Computational Intelligence in Data Mining (SHOCID), applying Computational Intelligence-paradigms, methods and techniques in the field of Data Mining, is being introduced. Currently available Data Mining... [more] Within this elaboration a generic system, subsequently referred to as System applying High Order Computational Intelligence in Data Mining (SHOCID), applying Computational Intelligence-paradigms, methods and techniques in the field of Data Mining, is being introduced. Currently available Data Mining systems are usually targeted on particular problem statements and require the user to understand how the underlying paradigms work, in contrary to the introduced one. SHOCID does not only fall back on complex Data Mining and Computational Intelligence techniques; it additionally does not require the user to understand how the result of a mining process is being achieved. Depending on the problem, the system is able to combine techniques and is, in some degree, able to decide on its own which strategy suits best. Within this elaboration known but adapted, as well as new approaches to Data Mining are being introduced, with focus on genericity and result-orientation for highlighting the aim of the research project: the provision of highly complex Computational Intelligence-techniques for mining data without the necessity of understanding these, implemented through a result-oriented interface and based on generic system architecture. The system's advantages are brought out by detailing one of its combinatorial data processing strategies as well as by describing algorithmically how training data for Feed Forward Artificial Neural Networks is synthesized. Finally, we provide an outline of the implemented techniques with focus on how the system makes use of them, always focusing on genericity.

• Problem-dependent, genetically evolving Data Mining solutions

  Florian Neukart, Sorin-Aurel Moraru, Peter Szakazs-Simon

  DAAAM International Vienna; 01/2011

  The already introduced generic Data Mining system SHOCID (System applying High Order Computational Intelligence in Data Mining) (Neukart et al., 2011) applies Computational Intelligence (CI) paradigms for solving any numeric Data Mining problem statement. Within this paper, we introduce the evolutio... [more] The already introduced generic Data Mining system SHOCID (System applying High Order Computational Intelligence in Data Mining) (Neukart et al., 2011) applies Computational Intelligence (CI) paradigms for solving any numeric Data Mining problem statement. Within this paper, we introduce the evolutionary approach by which the system is able to decide on its own, which of the possible evolutionary approaches suits best for solving a presented problem statement. Moreover, the system is, by the application of genetic algorithms, able to adapt the architecture and learning method of the Data Mining solution until coming to or at least close to the optimal solution.

• The use of Business Intelligence and Data Mining for the Improving of the Detection of Customer Needs in Service Engineering

  Helmut Aschbacher, Florian Neukart, Birgit Kammerhofer, Sebastian Schatzl

  SDN Conference; 01/2009

• Entwicklung eines prototypischen Vorgehensmodells mit webbasierter Dokumentenverwaltung für wissenschaftliche Konferenzen auf Basis eines Service Engineering Proceedings

  Florian Neukart

  01/2008

  Degree: Dipl.-Ing.

• Business Application Management Processes in an Exploration and Production Company

  Florian Neukart

  01/2007

  Degree: Dipl.-Ing. (FH)