Angel Barriga Barros

Autoregressive time series prediction by means of fuzzy inference systems using nonparametric residual variance estimation

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.

Automatic clustering-based identification of autoregressive fuzzy inference models for time series

We analyze the use of clustering methods for the automatic identification of fuzzy inference models for autoregressive prediction of time series. A methodology that combines fuzzy methods and residual variance estimation techniques is followed. A nonparametric residual variance estimator is used for a priori input and model selection. A simple scheme for initializing the widths of the input membership functions of fuzzy inference systems is proposed for the Improved Clustering for Function Approximation algorithm (ICFA), previously introduced for initializing RBF networks. This extension to the ICFA algorithm is shown to provide the most accurate predictions among a wide set of clustering algorithms. The method is applied to a diverse set of time series benchmarks. Its advantages in terms of accuracy and computational requirements are shown as compared to least-squares support vector machines (LS-SVM), the multilayer perceptron (MLP) and two variants of the extreme learning machine (ELM).

Enabling fuzzy technologies in high performance networking via an open FPGA-based development platform

Soft computing techniques and particularly fuzzy inference systems 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 networking equipment as well as other highly time-constrained application fields is however an open problem. We introduce a development platform for fuzzy inference systems with applications to network traffic analysis and control. The platform addresses the current requirements and constraints of high performance networking equipment. For the development process, we set up a methodology and a CAD tool chain that span the entire design process from initial specification in a high-level language to implementation on FPGA devices. An FPGA development board with PCI/PCIe interface is employed to support an open platform that comprises CAD tools as well as IP cores. PCI compatible fuzzy inference modules are implemented as System-on-Programmable-Chip (SoPC). We present satisfactory experimental results from the implementation of fuzzy systems for a number of applications in analysis and control of Internet traffic. These systems are shown to satisfy operational and architectural requirements of current and future high performance routing equipment. The platform proposed allows for the development of prototypes while avoiding large investments and complicated management procedures which constrain the testing and adoption of soft computing techniques in high performance networking.

Mining and Control of Network Traffic by Computational Intelligence

As other complex systems in social and natural sciences as well as in engineering, the Internet is hard to understand from a technical point of view. Packet switched networks defy analytical modeling. The Internet is an outstanding and challenging case because of its fast development, unparalleled heterogeneity and the inherent lack of measurement and monitoring mechanisms in its core conception.This monograph deals with applications of computational intelligence methods, with an emphasis on fuzzy techniques, to a number of current issues in measurement, analysis and control of traffic in the Internet. First, the core building blocks of Internet Science and other related networking aspects are introduced. Then, data mining and control problems are addressed. In the first class two issues are considered: predictive modeling of traffic load as well as summarization of traffic flow measurements. The second class, control, includes active queue management schemes for Internet routers as well as window based end-to-end rate and congestion control. The practical hardware implementation of some of the fuzzy inference systems proposed here is also addressed. While some theoretical developments are described, we favor extensive evaluation of models using real-world data by simulation and experiments.

Practical Course of Embedded Systems Based on XUPV2P Development Boards

This paper describes a lab course about embedded systems on field-programmable gate array. The proposed practices cover the main features of the design process, which include hardware architecture design, and embedded operating system configuration, adaptation, and implementation.

Modeling Time Series by Means of Fuzzy Inference Systems

In this chapter, we focus on long-term modeling and prediction of univariate nonlinear time series. First, a method for long-term time series prediction by means of fuzzy inference systems combined with residual variance estimation techniques is developed and validated through a number of time series prediction benchmarks. This method provides an automatic means of modeling and predicting network traffic load, and can thus be classified as a method for predictive data mining. Although the primary focus in this section is to develop a methodology for building simple and thus interpretable fuzzy inference systems, it will be shown that they also outperform some of the most accurate and commonly used techniques in the field of time series prediction.

Practical course of embedded Linux for XUPV2P development boards

This communication describes a lab course about embedded systems on FPGA. The proposed practices cover the main features of the design process, which includes the hardware architecture design, and the embedded operating system configuration, adaptation and implementation.

Summarization and Analysis of Network Traffic Flow Records

Current network measurement systems are becoming highly sophisticated, producing huge amounts of convoluted measurement data and statistics. As a very common case, those networks implementing statistics reporting based on the NetFlow [15] technology can generate several GBs of data on a daily basis. In addition, these measurements are often very hard to interpret. In this chapter we describe a method that provides linguistic summaries of network traffic measurements as well as a procedure for finding hidden facts in the form of linguistic association rules. Thus, here we address an association rules mining problem. The method is suitable for summarization and analysis of network measurements at the flow level. As a first step, fuzzy linguistic summaries are applied to analyze and extract concise and human consistent summaries from NetFlow collections. Then, a procedure for mining hidden facts in network flow measurements in the form of fuzzy association rules is developed. The method is applied to a wide set of heterogeneous flow measurements, and is shown to be of practical application to network operation and traffic engineering [6, 5], where it can help solve a number of current issues.

Predictive Models of Network Traffic Load

Understanding the dynamics and performance of packet switched networks on the basis of measurements enables practitioners to optimize resources. As network measurement research further advances and new measurement tools and infrastructures are available, the task of network operation becomes more and more complex. In this chapter we apply the methodology developed in the previous chapter to time series concerning network traffic load. An extensive predictability analysis is performed using the same nonparametric residual variance estimation technique that is integrated into the prediction methodology. Based on the predictability results, fuzzy inference based models that are both interpretable and accurate are derived for a wide set of heterogeneous time series for network traffic.

Open FPGA-Based Development Platform for Fuzzy Inference Systems

This chapter looks into the practical implementation of some of the fuzzy inference systems proposed in previous chapters. Both architectural and operational constraints are considered. The focus is on an open FPGA-based hardware platform for the implementation of efficient fuzzy inference systems for solving problems in high-performance packet switched networks. A feasibility study is conducted in order to show that the techniques developed can be deployed in current and future network scenarios with satisfactory performance.