Abstract: Kernel methods in learning machines have been developed in the last decade as new techniques for solving classification and regression problems. Kernel methods have many advantageous properties regarding their learning and generalization capabilities, but for getting the solution usually the computationally complex quadratic programming is required. To reduce computational complexity a lot of different versions have been developed. These versions apply different kernel functions, utilize the training data in different ways or apply different criterion functions. This paper deals with a special kernel network, which is based on the CMAC neural network. Cerebellar Model Articulation Controller (CMAC) has some attractive features: fast learning capability and the possibility of efficient digital hardware implementation. Besides these attractive features the modelling and generalization capabilities of a CMAC may be rather limited. The paper shows that kernel CMAC – an extended version of the classical CMAC network implemented in a kernel form – improves that properties of the classical version significantly. Both the modelling and the generalization capabilities are improved while the limited computational complexity is maintained. The paper shows the architecture of this network and presents the relation between the classical CMAC and the kernel networks. The operation of the proposed architecture is illustrated using some common benchmark problems.
Keywords: kernel networks, input-output system modelling, neural networks, CMAC, generalization error
Abstract: We present a tool to describe and simulate dynami systems. We use positive and negative influences. Our starting point is aggregation. We build positive and negative effects with proper transformations of the sigmoid function and using the conjunctive operator. From the input we calculate the output effect with the help of the aggregation operator. This algorithm is comparable with the concept of fuzzy cognitive maps.
Keywords: dynamic system, pliant concept, dombi operator
Abstract: Nowadays the competition among companies, joined to the environmental protection rules, is so compelling that they should not only be on the top of technology in they area, but also run their business according to life-long models. The emphasis on the product post-sale life is common for these models. The most popular model is Product Lifecycle Management, for manufacturing companies, or Service Engineering, for serviceoriented companies, and, for both, common paradigms are in maintenance, with conformance-to-use certification. The paper introduces basic research results achieved in application of Ambient Intelligence, and suggests considering maintenance as a cross section of the two business paradigms.
Keywords: Product Lifecycle Management, Service Engineering, Knowledge Based Systems, Condition Monitoring Maintenance, Ambient Intelligence
Abstract: The development of fuzzy control systems is usually performed by heuristic means, incorporating human skills, the drawback being in the lack of general-purpose development methods. A major problem, which follows from this development, is the analysis of the structural properties of the control system, such as stability, controllability and robustness. Here comes the first goal of the paper, to present a stability analysis method dedicated to fuzzy control systems with mechatronics applications based on the use of Popov’s hyperstability theory. The second goal of this paper is to perform the sensitivity analysis of fuzzy control systems with respect to the parametric variations of the controlled plant for a class of servo-systems used in mechatronics applications based on the construction of sensitivity models. The stability and sensitivity analysis methods provide useful information to the development of fuzzy control systems. The case studies concerning fuzzy controlled servo-systems, accompanied by digital simulation results and real-time experimental results, validate the presented methods.
Keywords: Mamdani fuzzy controllers, stability analysis, sensitivity analysis, mechatronics, servo-systems
Abstract: Frequent pattern mining is a heavily researched area in the field of data mining with wide range of applications. One of them is to use frequent pattern discovery methods in Web log data. Discovering hidden information from Web log data is called Web usage mining. The aim of discovering frequent patterns in Web log data is to obtain information about the navigational behavior of the users. This can be used for advertising purposes, for creating dynamic user profiles etc. In this paper three pattern mining approaches are investigated from the Web usage mining point of view. The different patterns in Web log mining are page sets, page sequences and page graphs.
Keywords: Pattern mining, Sequence mining, Graph Mining, Web log mining
Abstract: Fuzzy systems based on sparse rule bases produce the conclusion through approximation. This paper is the first part of a longer survey that aims to provide a qualitative view through the presentation of the basic ideas and characteristics of some methods and defining a general condition set brought together from an applicationoriented point of view.
Keywords: interpolative fuzzy reasoning, sparse rule base
Abstract: The paper proposes a new canonical form for RT-level descriptions, which can be systematically generated from both the specification and the structural description. The verification can be executed with the comparison of the two generated canonical form descriptions.
Abstract: Multiplicative operations for fuzzy numbers raise several problems both from the theoretical and practical point of view in fuzzy arithmetic. The multiplication based on Zadeh's extension principle and its triangular and trapezoidal approximation is used in several recent works in applications in geology. Recently, new product-type operation are introduced and studied, as e.g. the cross product of fuzzy numbers and the product obtained by the best trapezoidal approximation preserving the expeted interval. We present a comparative study of the above mentioned multiplications with respect to geological applications.
Keywords: fuzzy number, cross product, trapezoidal approximation, geological resource estimation
Abstract: This paper deals with the control of a dynamic system where the gains of the conventional PD controller are previously chosen by fuzzy methods in such a way as to obtain the optimal trajectory tracking. The gain factors are determined by solving fuzzy equations, and based on the sufficient possibility measure of the solution. It will be shown, that the rule premise for the given system input in fuzzy control system may also determine the possibility of realizing a rule. This possibility can be used for verifying the rule and for changing the rule-output, too. This leads to the optimization of the output. When calculating the possibility value the possible functional relation between the rule-premise and rule-consequence is taken into account. For defining the rule of inference in Fuzzy Logic Control (FLC) system special class of t-norm is used. The proposed fuzzy logic controller uses the functional relation between the rule premises and consequences, and the special class of pseudo-operators in the compositional rule of inference.
Keywords: FLC, fuzzification of the linear equations, PD controllers, pseudo-operators
Abstract: Cerebral blood flow (CBF) oscillation is a common feature of several physiological and pathophysiological states of the brains. In this study the characterization of the temporal pattern of the cerebral circulation has been analyzed. The classification of CBF signals has been carried out by two different classification methods – neural network and support vector machine – employing spectral and wavelet analysis as feature extraction techniques. The efficiency of these classification and feature extraction methods are evaluated and compared. Computations were carried out with Mathematica and its Wavelet as well as Neural Networks Applications.
Keywords: Biomedical Systems, Classification, Neural Network models, Radial base function networks, Support Vector Machine