Suchergebnisse

We propose an automatic methodology framework for short- and long-term prediction of time series by means of fuzzy inference systems. In this methodology, fuzzy techniques and statistical techniques for nonparametric residual variance estimation are combined in order to build autoregressive predictive models implemented as fuzzy inference systems. Nonparametric residual variance estimation plays a key role in driving the identification and learning procedures. Concrete criteria and procedures within the proposed methodology framework are applied to a number of time series prediction problems. The learn from examples method introduced by Wang and Mendel (W&M) is used for identification. The Levenberg–Marquardt (L–M) optimization method is then applied for tuning. The W&M method produces compact and potentially accurate inference systems when applied after a proper variable selection stage. The L–M method yields the best compromise between accuracy and interpretability of results, among a set of alternatives. Delta test based residual variance estimations are used in order to select the best subset of inputs to the fuzzy inference systems as well as the number of linguistic labels for the inputs. Experiments on a diverse set of time series prediction benchmarks are compared against least-squares support vector machines (LS-SVM), optimally pruned extreme learning machine (OP-ELM), and k-NN based autoregressors. The advantages of the proposed methodology are shown in terms of linguistic interpretability, generalization capability and computational cost. Furthermore, fuzzy models are shown to be consistently more accurate for prediction in the case of time series coming from real-world applications. ; The first author is supported by a Marie Curie Intra-European Fellowship for Career Development (grant agreement PIEF-GA-2009-237450) within the European Community´s Seventh Framework Programme (FP7/20072013). This work has been supported in part by project TEC2008-04920/MICINN from the Spanish Ministry of Science and Innovation, as well as project P08-TIC-03674 and grants IAC07-I-0205:33080 and IAC08-II-3347:56263 from the Andalusian regional Government. ; Peer reviewed

Trabajo presentado al "FUZZ-IEEE 2008" celebrado en Hong Kong del 1 al 6 de Junio de 2008. ; We apply fuzzy techniques for system identification and supervised learning in order to develop fuzzy inference based autoregressors for time series prediction. An automatic methodology framework that combines fuzzy techniques and statistical techniques for nonparametric residual variance estimation is proposed. Identification is performed through the learn from examples method introduced by Wang and Mendel, while the Marquard-Levenberg supervised learning algorithm is then applied for tuning. Delta test residual noise estimation is used in order to select the best subset of inputs as well as the number of linguistic labels for the inputs. Experimental results for three time series prediction benchmarks are compared against LS-SVM based autoregressors and show the advantages of the proposed methodology in terms of approximation accuracy, generalization capability and linguistic interpretability. ; This work has been supported in part by project TEC2005-04359/MIC from the Spanish Ministry of Education and Science as well as project TIC2006-635 and grant IAC07-I-0205:33080 from the Andalusian regional Government. ; Peer Reviewed

Prólogo del libro: "Mining and Control of Network Traffic by Computational Intelligence". ; The first author is supported by a Marie Curie Intra-European Fellowship for Career Development (grant agreement PIEF-GA-2009-237450) within the European Community´s Seventh Framework Programme (FP7/20072013). Most of this work was done while the first author was with the Microelectronics Institute of Seville, IMSECNM, CSIC. This work was supported in part by the European Community under the MOBY-DIC Project FP7-IST-248858 (www.mobydic-project.eu). The research presented here has been supported in part by a PhD studentship from the Andalusian regional Government, project TEC2008-04920, from the Spanish Ministry of Education and Science, as well as project P08-TIC-03674 from the Andalusian regional Government. ; Peer reviewed

Trabajo presentado al "FUZZ-IEEE 2008" celebrado en Hong Kong del 1 al 6 de Junio de 2008. ; We address, by means of fuzzy linguistic summaries, two related problems: summarizing network flow statistics and making these statistics human-readable. Two complementary summarization methods are developed. First, a fixed set of protoforms of interest is defined, and the ones with a higher truth value are shown to the user as simple on-line summaries. This first method is suitable for real-time monitoring. Then, an association rules mining process is carried out in order to find hidden relations in flow records. Both approaches are implemented in a tool capable of real-time and off-line processing of network flow records. Experimental results for a number of heterogeneous NetFlow records show the usefulness of linguistic summaries to both network practitioners and users. ; This work has been supported in part by projects TEC2005-04359/MIC from the Spanish Ministry of Education and Science and project TIC2006-635 from the Andalusian regional Government. ; Peer Reviewed

Comunicación presentada al "DCIS 2007" celebrada en Sevilla del 21 al 23 de Noviembre de 2007. ; Soft computing techniques are gaining momentum as tools for network traffic modeling, analysis and control. Efficient hardware implementations of these techniques that can achieve real-time operation in high-speed communications equipment is however an open problem. This paper describes a platform for the development of fuzzy systems with applications to communications systems, namely network traffic analysis and control. An FPGA development board with PCI interface is employed to support an open platform that comprises open CAD tools as well as IP cores. For the development process, we set up a methodology and a CAD tools chain that cover from initial specification in a high-level language to implementation on FPGA devices. PCI compatible fuzzy inference modules are implemented as SoPC based on the open WISHBONE interconnection architecture. We outline results from the design and implementation of fuzzy analyzers and regulators for network traffic. These systems are shown to satisfy operational and architectural requirements of current and future high-performance routing equipment. ; This work has been supported in part by projects TEC2005-04359/MIC from the Spanish Ministry of Education and Science and TIC2006-635 from the Andalusian regional Government. ; Peer Reviewed

Comunicación presentada a la "11th International Conference Information Processing and Management of Uncertainty in Knowledge-based Systems" celebrada en Paris del 2 al 7 de Julio de 2006. ; A major research problem in Internet transport and network layers is the development of traffic regulation mechanisms that can cope with the requirements of a growing diversity of technologies, applications and services. This paper presents novel mechanisms for intelligent traffic scheduling in Internet routers by means of fuzzy logic based systems. A systematic design methodology, interpretability principles, evaluation over a broad range of network scenarios as well as practical implementation constraints have been considered. A comparative evaluation of results obtained by means of our fuzzy controllers as compared to that of traditional approaches is outlined. ; This work has been partially supported by projects TEC2005-04359/MIC from the Spanish Ministry of Education and Science as well as TIC2006-635 from the Andalusian Regional Government. ; Peer Reviewed

Comunicación presentada a la "IEEE International Conference on Fuzzy Systems" celebrada en Vancouver (Canadá) del 16 al 21 de Julio de 2006. ; End-to-end Internet packet dynamics is a complex problem for which models available to date are at best incomplete. A major research problem in Internet transport layer protocols is the development of rate control mechanisms that can cope with the requirements of a growing diversity of technologies, applications and services. This paper describes novel mechanisms for intelligent end-to-end traffic rate control in Internet by means of fuzzy systems. We first outline a fuzzy logic based generalization of TCP (Transport Control Protocol) rate control principles. The design of a fuzzy TCP-like window-based rate controller is then described. A systematic fuzzy systems design methodology is used in order to simulate and implement the system as an experimental tool. A comparative evaluation of simulation and implementation results from the fuzzy rate controller as compared to that of traditional controllers is outlined. Besides being a useful modelling approach, the fuzzy rule based rate controller is shown to outperform other approaches with regards to a number of criteria. ; This work has been supported in part by projects TEC2005-04359/MIC from the Spanish Ministry of Education and Science and TIC2006-635 from the Andalusian regional Government. ; Peer Reviewed

Comunicación presentada a la "11th International Conference Information Processing and Management of Uncertainty in Knowledge-based Systems" celebrada en Paris del 2 al 7 de Julio de 2006. ; The characteristics of the new version of the fuzzy systems development environment Xfuzzy is presented. The environment covers the aspects related to the specification, verification, adjustment and implementation of fuzzy systems. It is an open environment (in the sense that the user can define many functional and structural aspects) and a free distribution tool that allows proving new formalisms and helps the definition and implementation of complex systems. ; This work was partially supported projects TEC2005-04359/MIC from the Spanish Ministry of Education and Science as well as TIC2006-635 from the Andalusian Regional Government. ; Peer Reviewed