Abstract:An attribute can be Boolean or quantitative. There are lots of systems and methods for mining Boolean association rules but few for quantitative. Mapping quantitative attributes into Boolean attributes is a convenient and efficient way. In this paper, a new clustering algorithm is presented. Quantitative attribute values are partitioned into intervals according to the distribution of them in database. Then the intervals are mapped into Boolean attributes. In this way, quantitative rules can be mined by using the techniques of mining Boolean association rules.

Abstract:In this paper, based on MPEG-2 Test Model 5 (TM5) bit rate control algorithm, there are some improvements proposed: ① an intelligent integral action is proposed for improving the control precision of bit rate; ② after using DCT AC coefficients to define the measure of macroblock and picture coding difficulty, an improvement on target bits allocation is proposed for the decoded picture quality and conformance including scene change, which can adjust the target bits allocation adaptively by their measure. Simulation results show that it can achieve high control precision by using intelligent integral action, its relative error is reduced one order of magnitude compared with TM5; using the improvement on target bits allocation, the average decoding PSNR is increased visibly, and the decoded picture quality is improved when scene changes.

Abstract:To take the merits of both wireframe and shaded surfaces, sometimes it is advantageous to combine them together in a single image. However, the combination would often lead to edge stitching and might draw edges that are hidden and the current methods have not so far completely solved the problems since they are mostly based on the Z-buffer comparison. A new method without Z-buffer comparison is presented in this paper. By this method, overlay of edges on shaded surfaces is realized by employment of the correspondence between edges and polygons as well as the hidden relation between polygons. Thus, the existing problems for wireframe overlaying shaded surfaces could be removed, and the calculation for depth values of edges may be saved.

Abstract:In this paper, a new higher-order bidirectional associative memory model is presented. It is an extension of Tai's HOBAM(higher-order bidirectional associative memory) and Jeng's MIBAM(modified intraconnected BAM). The stability of the new model, in synchronous and asynchronous updating modes, is proven by defining an energy function such that it can ensure all the training pattern pairs to become its asymptotically stable points. Using statistical analysis principle, the storage capacity of the proposed model is estimated. The computer simulations show that this model has not only higher storage capacity but also better error-correcting capability.

Abstract:Based on the comparison of the conventional definitions of B-tree and B⁺-tree, and their manipulating algorithms, a new definition of B⁺-tree: RFN-B⁺-tree is presented in order to achieve higher space efficiency and higher availability. Its space efficiency compared with conventional B⁺-tree is analyzed firstly. Then the availability of RFN-B⁺-tree index file in terms of its full-link pointer structure together with two modules that support this availability: random access algorithm based on virtual root node, and node restructuring algorithm are discussed.

Abstract:There exists a very interesting feature in SAT(satisfiability) problem. With a fixed numbers of variables, the satifactory probability of an SAT instance change sharply from 1 to 0 while the number of clauses increasing, and the phase transition point is estimated to be K≈4.3*N. The phase transitions are of great importance to the efficient algorithms designing to solve the SAT problem. More generally, the phase transition of 2-3-SAT problem was discussed in this paper. The analysis of the location of the phase transition point of 2-3-SAT shows that there is an linear ratio between a 2-clause and a 3-clause in the sense of the constraint power which could help to design a more powerful heuristic for algorithms designing.

Abstract:Generalize the methods of paraconsistent fuzzy reasoning into modal logic, propose a paraconsistent modal logic, whose logical consequence is a modal extension of a paraconsistent fuzzy implication, which has the abilities of both handling inconsistency and representing multi-world modals, and present its sound and complete Gentzen style inference system.

Abstract:Paraconsistent circumscription LPc is a non-classical logic which has both nonmonotonicity and paraconsistency. It can be viewed as a formalism of commonsense reasoning with incomplete and inconsistent knowledge. In this paper, the results of computational complexity on propositional LPc and its implementation algorithm is presented. It is showed that LPc is NP-complete. A linear algorithm to transform LPc into equivalent prioritized circumscription is provided. Because there are a number of useful algorithms for circumscription which could be implemented by resolution, a new way to implementation of LPc is drawn.

Abstract:It is an important and hard technique problem in the area of color printing to keep the accordance of color between the different devices. In this paper, from the view of machine learning, a new color space transformation method based on machine learning is presented. Using the basic ideas of science discovery, it successfully solves the automatic and general problem in the different color ink printers. The experimental results show that this approach is superior to the traditional methods. Currently, this technique has been applied to the new products in the color ink printer markets, under the cooperation with Canotec Corporation Incorporated in Japan.

Abstract:After the modern models for network computing are analyzed, their weak spots of the representation of serving resources are pointed. And a new interface model which can specify a computing resource more precisely and comprehensively is designed in order to improve the interaction of computers themselves so that the software resources can be searched and shared automatically.

Abstract:In this paper, the algorithm to the truth value and information amount of compound fuzzy proposition from fuzzy information amount is presented. The new algorithm is effective because it is relative to logic and relative importance between sub-propositions, and loses a little information. Through the experiments, the new algorithm eliminates the drawback of losing information, closing and invalidates of whole equality and ambiguity, and the reasonable results can be obtained.

Abstract:The main idea of variant design is to adapt the existing product or module to the new requirements. Based on the analysis of the basic problem solving model of variant design, a case-based reasoning system combined with rule-based reasoning and constraint satisfaction method to imitate the process of variant design is proposed in this paper. With the application——“spindle machine module variant design system” as an example, the knowledge representation model and the reasoning process of the variant-design-oriented cooperative reasoning system are discussed in detail.

Abstract:The authors promote the concept of software architecture engineering, establish the life cycle of software architecture, classify the different kinds of software architectures and architecture modeling, introduce several typical software architecture modeling languages, give a formal architectural style modeling language XYZ/SAE(as a sub-language of a series of temporl logic languages——XYZ/E) in this paper. It supports the description of component, connector and configuration, and also static and dynamic aspects of software architecture.

Abstract:FP algorithm is improved in this paper. First, the “common characteristic bit” is optimized, then the selection of x⁰ sample is optimized. Finally, the design of threshold is optimized. After these optimizations, the attractive domain of sample appears more even. The solution is proved by the test data.

Abstract:Deeply analyzed the conventional genetic algorithm and for its shortcomings on numerical optimization, evolution genetic algorithm (EGA) is proposed. EGA makes some improvements on description of genes of chromosome, genetic operators of crossover and mutation, designing of fitness function, selection method on chromosome of candidate solutions and convergence criterion. The optimization results of some functions (including hard Shekel' function) show that EGA has fine ability of global searching and speedy convergence.