Abstract
This paper presents a new approach to perform on-line dynamic security assessment and monitoring of electric power systems exploiting a statistical hybrid learning technique - the Kernel Regression Trees This technique, besides producing fast;security classification, can still quantify, hi real-time, the security degree of the system, by emulating continuos security indices that translate the power system dynamic behavior. Moreover it can provide interpretable security structures. The feasibility of this approach was demonstrated in the dynamic security assessment of isolated systems with large amounts of wind power production, like In the Crete island electric network (Greece) Comparative results regarding performances of Decision Trees and Neural Networks are also presented and discussed. From the obtained results, the proposed approach showed to provide good predicting structures whose performance stands up to the performance of the two other existent methods.

Abstract
This paper presents a new strict, purely functional implementation of attribute grammars. Incremental evaluation is obtained via standard function memoization. Our new implementation of attribute grammars increases the incremental behaviour of the evaluators by both reducing the memoization overhead and increasing their potential incrementallity. We present also an attribute grammar transformation, which increases the incremental performance of the attribute evaluators after a change that propagates its effects to all parts of the syntax tree. These techniques have been implemented in a purely functional attribute grammar system and the first experimental results are presented.

Abstract
It is shown here that st straightforward procedure can be used to quantize the linearized equations for an electromagnetic field in a plasma. This leads to a definition of an effective mass for the transverse photons, and a different one far the longitudinal photons, or plasmons. Both masses are simply proportional to the electron plasma density. A nonlinear perturbative analysis can also be used to extend the quantization procedure, in order to include the ponderomotive force effects. This leads to the definition of a photon charge operator. The mean value of this operator, for a quantum state with a photon occupation number equal to 1, is the equivalent charge of the photon in a plasma.

Abstract
In this paper, we present a new algorithm for motion flow field estimation using Cellular Neural Networks (CNN). We start from a mathematical viewpoint (i.e. statistical regularisation based on Markov Random Field (MRF) and proceed by mapping the algorithm onto a cellular neural network. Because of the temporal dynamics inherent in the cells of the CNN it is well suited to processing time-varying images. A robust motion estimation algorithm is achieved by using a spatio-temporal neighbourhood for modelling pixel interactions.

Abstract
Geographical Information Systems were originally intended to deal with snapshots representing a single state of some reality but there are more and more applications requiring the representation and querying of time-varying information. This work addresses the representation of moving objects on GIS. The continuous nature of movement raises problems for representation in information systems due to the limited capacity of storage systems and the inherently discrete nature of measurement instruments. The stored information has therefore to be partial and does not allow an exact inference of the real-world object's behavior. To cope with this, query operations must take uncertainty into consideration in their semantics in order to give accurate answers to the users. The paper proposes a set of operations to be included in a GIS or a spatial database to make it able to answer queries on the spatio-temporal behavior of moving objects. The operations have been selected according to the requirements of real applications and their semantics with respect to uncertainty is specified. A collection of examples from a case study is included to illustrate the expressiveness of the proposed operations.

Abstract
This paper describes an approach to multivariate regression that aims at improving the computational efficiency and comprehensibility of local regression techniques. Local regression modeling is known for its ability to accurately approximate quite diverse regression surfaces with high accuracy. However, theses methods are also known for being computationally demanding and for not providing any comprehensible model of the data. These two characteristics can be regarded as major drawbacks in the context of a typical data, mining scenario. The method we describe tackles these problems by integrating local regression within a partition-based induction method.