Vol. 32 · No. 5

This research presents inverted lists as a new data structure for the dynamic dictionary matching algorithm. The inverted lists structure, which derives from the inverted index, is implemented by the perfect hashing table. The dictionary is constructed in optimal time and the individual patterns can be updated in minimal time. The searching phase scans the given text in a single pass, even in a worst case scenario. In experimental results, the inverted lists used less time and space than the traditional structures; the searches were processed and showed an efficient linear time.

Argumentation frameworks for ontology reasoning and management have received extensive interests in the field of artificial intelligence in recent years. As one of the most popular argumentation frameworks, Besnard and Hunter's framework is built on arguments in the form of łangle Phi, phi rangle where Phi is consistent and minimal for entailing phi. However, the problem of generating arguments over ontologies is still open. This paper presents an approach to generating arguments over DL-Lite ontologies by searching support paths in focal graphs. Moreover, theoretical results and examples are provided to ensure the correctness of this approach. Finally, we show that approach has the same complexity as propositional revision.

The paper focuses on the process of selecting representative sample documents written in a natural language that can be used as the basis for automatic selection or classification of textual documents. A method of selecting the examples from a larger set of candidate examples, called automatic biased sample selection, is compared to random and manual selection. The methods are evaluated by experiments carried out with real world data consisting of customer reviews, with different document representations and similarity measures. The presented approach, that provided satisfactory results, faces problems related to processing user created content and huge computational complexity and can be used as an alternative to manual selection and evaluation of textual samples.

For Web services composition problem, this paper proposes an executable code creation algorithm to model Web services composition based on Petri net model, and develops an executable composition code automatic creation tool. This tool can achieve the automatic creation process from composition model to executable code, and more meaningfully makes it possible to analyze and validate composition process logically. Finally, experiment results have proven that the tool of this paper is feasible.

Data uncertainty inherently exists in a large number of applications due to factors such as limitations of measuring equipments, update delay, and network bandwidth. Recently, modeling and querying uncertain data have attracted considerable attention from the database community. However, how to perform advanced analysis on uncertain data remains an interesting question. In this paper, we focus on the execution of skyline computation over uncertain moving objects. We propose a novel probabilistic skyline model where an uncertain object may take a probability to be in the skyline at a certain time point, therefore a p-t-skyline contains those moving objects whose skyline probabilities are at least p at time point t. Computing probabilistic skyline over a large number of uncertain moving objects is a daunting task in practice. In order to efficiently compute the probabilistic skyline query, we propose a discrete-and-conquer strategy, which follows the sampling-bounding-pruning-refining procedure. To further reduce the skyline computation cost, we propose an enhanced framework that is based on a multi-dimensional indexing structure combined with the discrete-and-conquer strategy. Through extensive experiments with synthetic datasets, we show that the framework can efficiently support skyline queries over uncertain moving object and is scalable on large data sets.

Visual methods have been extensively studied and performed in cluster data analysis. Given a pairwise dissimilarity matrix D of a set of n objects, visual methods such as Enhanced-Visual Assessment Tendency (E-VAT) algorithm generally represent D as an n times n image I( overlineD) where the objects are reordered to expose the hidden cluster structure as dark blocks along the diagonal of the image. A major constraint of such methods is their lack of ability to highlight cluster structure when D contains composite shaped datasets. This paper addresses this limitation by proposing an enhanced visual analysis method for cluster tendency assessment, where D is mapped to D' by graph based analysis and then reordered to overlineD' using E-VAT resulting graph based Enhanced Visual Assessment Tendency (GE-VAT). An Enhanced Dark Block Extraction (E-DBE) for automatic determination of the number of clusters in I( overlineD') is then proposed as well as a visual data partitioning method for cluster formation from I( overlineD') based on the disparity between diagonal and off-diagonal blocks using permuted indices of GE-VAT. Cluster validation measures are also performed to evaluate the cluster formation. Extensive experimental results on several complex synthetic, UCI and large real-world data sets are analyzed to validate our algorithm.

The creation of an initial glossary of terms is a preliminary phase of domain ontology building. Existing methodologies assume that such a glossary has been created by analysing existing documents or using expert knowledge. Some methods have been defined for this step of ontology building; these methods are mostly based on the analysis of existing documents. We propose to utilise the existing pieces of knowledge obtained in the area of object-oriented analysis; the description of a domain structure, behaviour and rules. Domain structure, behaviour and rules all together represent a complex and systematic view of the domain that makes it possible to create a high-quality glossary. This method is demonstrated using the domain of a road traffic system. Our method has been developed as an extension of the well-known METHONTOLOGY method. Our extension is general enough to be relevant for other ontology-building methodologies.

This paper presents a study on a rule induction application for generating an agent strategy. It is a new approach in multi-agent systems, where reinforcement learning and evolutionary computation is broadly used for this purpose. Experimental results show that rule induction improves agent performance very quickly. What is more, rule-based knowledge representation has many advantages. It is comprehensive and clear. It allows for the examination of the learned knowledge by humans. Because of modularity of the knowledge, it also allows for the implementation of the knowledge exchange in a natural way – only necessary set of rules can be sent. Rule induction is tested in two domains: Fish Banks game, in which agents run fishing companies and learn how to allocate ships, and Predator-Prey domain, in which predator agents learn how to capture preys. The proposed learning mechanism should be beneficial in all domains, in which agents can determine the results of their actions.

In the data mining field, data representation turns out to be one of the major factors affecting mining algorithm scalability. Mining Frequent Itemsets (MFI) is a data mining problem that is heavily affected by this fact. The vertical approach is one of the successful data representations adopted for MFI problem. The main advantage of this approach is support for fast frequency counting via joining operations. Recently, an encoding method called prime-encoding is proposed as an enhancement for the vertical approach citeKaram10:JCSS. The performance study introduced in citeKaram10:JCSS confirmed the high quality of prime-encoding based vertical mining of frequent sequence over other vertical and horizontal ones in terms of space and time. Though sequence mining is more general than itemset mining, this paper presents a prime-encoding based vertical mining of frequent itemsets with new optimizations and a new re-encoding method that further enhance memory and speed. The experimental results show that prime encoding based vertical itemset mining is suitable for high-dimensional sparse data.

A Science Gateway is a connection between scientists and their computational tools in the form of web portal. It creates a space for communities, collaboration and data sharing and visualization in a comprehensive and efficient manner. The main purpose of such a solution is to allow users to access the computational resources, process and analyze their data and get the results in a uniform and user friendly way. In this paper we propose a complex solution based on the Rich Internet Application (RIA) approach consisting of a web portal powered by Vine Toolkit with Adobe Flex/BlazeDs technologies. There are two Science Gateways described in detail one for engineers to manage computationally intensive workflows used in advanced airplane construction simulations, and one for nanotechnology scientists to manage experiments in nano-science field calculated with Density Functional Theory (DFT). In both cases the results show how modern web solution can help scientists in their work.