基于类-属性关联度的启发式离散化技术

连续属性离散化在数据挖掘、机器学习和人工智能等领域起着重要的作用.鉴于此,提出一种基于类-属性关联度的启发式离散化技术.该技术定义了一个新的离散化标准,根据数据本身的特性选择最佳断点,克服了目前最先进自顶向下离散化方法存在的缺陷.基于粗糙集理论中变精度粗糙集模型,提出一种新的不一致衡量标准,能够有效地控制离散化所产生的信息丢失,允许数据存在适当的分类错误度.实验结果和统计性分析表明,所提出的技术显著地提高了J4.8决策树和SVM分类器的学习精度.     

The universal remote console: a universal access bus for pervasive computing

V2, a technical committee of the International Committee for information technology standards is developing standards for information technology access interfaces. As part of this endeavor, V2 is to issue the Universal Remote Console architecture, a set of draft standards for public review and comment. The five standards specify communications between a Target - a device or service the user wishes to access and a universal remote console-software that is typically hosted on a user's personal device. The standards specify mechanisms for the discovery, selection, configuration, and operation of user interfaces and its options.     

Multiresolution state-space discretization for Q-Learning with pseudorandomized discretization

A multiresolution state-space discretization method with pseudorandom gridding is developed for the episodic unsupervised learning method of Q-learning.It is used as the learning agent for closed-loop control of morphing or highly reconfigurable systems.This paper develops a method whereby a state-space is adaptively discretized by progressively finer pseudorandom grids around the regions of interest within the state or learning space in an effort to break the Curse of Dimensionality.Utility of the method is demonstrated with application to the problem of a morphing airfoil,which is simulated by a computationally intensive computational fiuid dynamics model.By setting the multiresolution method to define the region of interest by the goal the agent seeks,it is shown that this method with the pseudorandom grid can learn a specific goal within ±0.001 while reducing the total number of state-action pairs needed to achieve this level of specificity to less than 3000.     

Controller discretization: a gap metric framework for analysis and synthesis

Although techniques for directly synthesising sampled-data (SD) compensators are available in the literature, feedback controller design is perhaps best understood in a purely continuous-time setting. As such, a feedback controller is often designed in the continuous-time domain and then discretized for digital implementation. It is important for the discretization step involved to yield an SD approximation which captures the essential features of the original controller from the perspective of closed-loop behavior. In this note, a gap metric framework is developed for studying the controller discretization problem for linear time-invariant (LTI) plants and controllers. Importantly, knowledge of a gap metric distance between an LTI controller and an SD approximation permits explicit characterization of the possible difference in closed-loop performance, with any LTI plant for which the LTI controller is known to work well, accounting for intersample behavior. The central result of the new framework gives rise to an algorithm for computing a gap metric measure of the distance between an LTI controller and a given discretization, and a technique for synthesising a SD approximation which is optimal with respect to this metric.     

Optimal control in non-convex domains: a priori discretization error estimates

Abstract An optimal control problem for a two-dimensional elliptic equation with pointwise control constraints is investigated. The domain is assumed to be polygonal but non-convex. The corner singularities are treated by a priori mesh grading. Approximations of the optimal solution of the continuous optimal control problem are constructed by a projection of the discrete adjoint state. It is proved that these approximations have convergence order h². Keywords Linear-quadratic optimal control problems, error estimates, elliptic equations, non-convex domains, corner singularities, control constraints, superconvergence. Mathematics Subject Classification (2000): 49K20, 49M25, 65N30, 65N50     

On the capacitated concentrator location problem: a reformulation by discretization

In this paper, we present and discuss a discretized model for the two versions of the capacitated concentrator location problem: a simple version (SCCLP) and a version with modular capacities (MCCLP). We show that the linear programming relaxation of the discretized model is at least as good as the linear programming relaxation of conventional models for the two variations of the problem under study. A technique for deriving valid inequalities from the equations of the discretized model is also given. We will show that this technique provides inequalities that significantly enhance the linear programming bound of the discretized model. Our computational results show the advantage of the new models for obtaining the optimal integer solution for the two versions of the problem.     

Simulation of a magneto-mechanical damping machine: analysis,discretization, results

This paper presents and analyzes a method for the simulation of the dynamical behavior of a coupled magneto-mechanical system such as a damping machine. We consider a two-dimensional model based on the transverse magnetic formulation of the eddy currents problem for the electromagnetic part and on the motion equation of a rotating rigid body for the mechanical part.The magnetic system is discretized in space by means of Lagrangian finite elements and the sliding mesh mortar method is used to account for the rotation. In time, a one step Euler method is used, implicit for the magnetic and velocity equations. The coupled differential system is solved with an explicit procedure.     

CAIM discretization algorithm

The task of extracting knowledge from databases is quite often performed by machine learning algorithms. The majority of these algorithms can be applied only to data described by discrete numerical or nominal attributes (features). In the case of continuous attributes, there is a need for a discretization algorithm that transforms continuous attributes into discrete ones. We describe such an algorithm, called CAIM (class-attribute interdependence maximization), which is designed to work with supervised data. The goal of the CAIM algorithm is to maximize the class-attribute interdependence and to generate a (possibly) minimal number of discrete intervals. The algorithm does not require the user to predefine the number of intervals, as opposed to some other discretization algorithms. The tests performed using CAIM and six other state-of-the-art discretization algorithms show that discrete attributes generated by the CAIM algorithm almost always have the lowest number of intervals and the highest class-attribute interdependency. Two machine learning algorithms, the CLIP4 rule algorithm and the decision tree algorithm, are used to generate classification rules from data discretized by CAIM. For both the CLIP4 and decision tree algorithms, the accuracy of the generated rules is higher and the number of the rules is lower for data discretized using the CAIM algorithm when compared to data discretized using six other discretization algorithms. The highest classification accuracy was achieved for data sets discretized with the CAIM algorithm, as compared with the other six algorithms.     

A discretization algorithm based on a heterogeneity criterion

Discretization, as a preprocessing step for data mining, is a process of converting the continuous attributes of a data set into discrete ones so that they can be treated as the nominal features by machine learning algorithms. Those various discretization methods, that use entropy-based criteria, form a large class of algorithm. However, as a measure of class homogeneity, entropy cannot always accurately reflect the degree of class homogeneity of an interval. Therefore, in this paper, we propose a new measure of class heterogeneity of intervals from the viewpoint of class probability itself. Based on the definition of heterogeneity, we present a new criterion to evaluate a discretization scheme and analyze its property theoretically. Also, a heuristic method is proposed to find the approximate optimal discretization scheme. Finally, our method is compared, in terms of predictive error rate and tree size, with Ent-MDLC, a representative entropy-based discretization method well-known for its good performance. Our method is shown to produce better results than those of Ent-MDLC, although the improvement is not significant. It can be a good alternative to entropy-based discretization methods.     

Effects of constructing fuzzy discretization from crisp discretization for rule-based classifiers

Crisp discretization is one of the most widely used methods for handling continuous attributes. In crisp discretization, each attribute is split into several intervals and handled as discrete numbers. Although crisp discretization is a convenient tool, it is not appropriate in some situations (e.g., when there is no clear boundary and we cannot set a clear threshold). To address such a problem, several discretizations with fuzzy sets have been proposed. In this paper we examine the effect of fuzzy discretization derived from crisp discretization. The fuzziness of fuzzy discretization is controlled by a fuzzification grade F. We examine two procedures for the setting of F. In one procedure, we set F beforehand and do not change it through training rule-based classifiers. In the other procedure, first we set F and then change it after training. Through computational experiments, we show that the accuracy of rule-based classifiers is improved by an appropriate setting of the grade of fuzzification. Moreover, we show that increasing the grade of fuzzification after training classifiers can often improve generalization ability. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

Passive image forensics using universal techniques: a review

Digital tamper detection is a substantial research area of image analysis that identifies the manipulation in the image. This domain has matured with time and incredible accuracy in the last five years using machine learning and deep learning-based approaches. Now, it is time for the evolution of fusion and reinforcement-based learning techniques. Nevertheless, before commencing any experimentation, a researcher needs a comprehensive state of the art in that domain. Various directions, their outcome, and analysis form the basis for successful experiments and ensure better results. Universal image forensics approaches are a significant subset of image forensic techniques and must be explored thoroughly before experimentation. This motivated authors to write a review of these approaches. In contrast to the existing recent surveys that aim at image splicing or copy-move detection, our study aims to explore the universal type-independent techniques required to highlight image tampering. Several universal approaches based on resampling, compression, and inconsistency-based detection are compared and evaluated in the presented work. This review communicates the approach used for review, analysed literature, and lastly, the conclusive remarks. Various resources beneficial for the research community, i.e. journals and datasets, are explored and enumerated. Lastly, a futuristic reinforcement learning-based model is proposed.

     

连续属性的频数监督断点离散化技术

提出了一种频数监督断点的离散化算法。该算法利用所提出的频数监督断点思想产生初始断点,并在此基础上进行断点简约。实验结果表明该算法所产生的断点不仅符合实际数据分布,而且更为合理、精练。     

模糊系统万能逼近理论研究综述

模糊系统通用逼近理论是模糊理论研究的一个重要方向．目前，对模糊系统通用逼近性的研究已经取得了很大的进展．对模糊系统的通用逼近性、模糊系统作为通用逼近器的充分条件和必要条件以及模糊系统的逼近精度等方面的研究进行了较为详尽的综述，分析了各种分析方法的主要成果及其特点（包括优点和局限性），并指出了今后模糊系统通用逼近理论研究中有待解决的许多问题．     

XML: data's universal language

You have probably heard a lot lately about Extensible Markup Language (XML): a new tag based language for describing data. Trade journals and industry magazines devote cover pages to it, and application vendors build in support for it and promote its virtues. The most enthusiastic supporters claim XML will open the door for the next generation of Internet applications. But will XML live up to all the hype? XML is fast becoming the key language for an increasing number of new applications. It is poised to fundamentally alter the way we deliver and use information and to enable the creation of new, powerful applications. It has the potential to shape the future of the Web, among other things. But XML is still a work in progress. Designers and developers are adding new features to it and developing new technologies around it