MINING OF COMPLEX OBJECTS VIA DESCRIPTION CLUSTERING

Alejandro García López, Rafael Berlanga and Roxana Danger

In this work we intend to present a formal framework for mining complex objects, being those characterised by a set of attributes and their corresponding values. First we will do an introduction of the various Data Mining techniques available in the literature to extract association rules. We will as well show some of the drawbacks of these techniques and how our proposed solution is going to tackle them. Then we will show how applying a clustering algorithm as a pre-processing step on the data allow us to find groups of attributes and objects that will provide us with a richer starting point for the Data Mining process. Then we will define the formal framework, its decission functions and its interesting measurement rules, as well as a newly designed Data Mining algorithms that will be specifically tuned for our objective. We will also show the type of knowledge to be extracted in the form of a set of association rules. Finally we will state our conclusions and propose the future work.

APPROXIMATE REASONING TO LEARN CLASSIFICATION RULES

Amel Borgi

In this paper, we propose an original use of approximate reasoning not only as a mode of inference but also as a means to refine a learning process. This work is done within the framework of the supervised learning method SUCRAGE which is based on automatic generation of classification rules. Production rules whose conclusions are accompanied by belief degrees, are obtained by supervised learning from a training set. These rules are then exploited by a basic inference engine: it fires only the rules with which the new observation to classify matches exactly. To introduce more flexibility, this engine was extended to an approximate inference which allows to fire rules not too far from the new observation. In this paper, we propose to use approximate reasoning to generate new rules with widened premises: thus imprecision of the observations are taken into account and problems due to the discretization of continuous attributes are eased. The objective is then to exploit the new base of rules by a basic inference engine, easier to interpret. The proposed method was implemented and experimental tests were carried out.

A PATTERN SELECTION ALGORITHM IN KERNEL PCA APPLICATIONS

Ruixin Yang, John Tan and Menas Kafatos

Principal Component Analysis (PCA) has been extensively used in different fields including earth science for spatial pattern identification. However, the intrinsic linear feature associated with standard PCA prevents scientists from detecting nonlinear structures. Kernel-based principal component analysis (KPCA), a recently emerging technique, provides a new approach for exploring and identifying nonlinear patterns in scientific data. In this paper, we recast KPCA in the commonly used PCA notation for earth science communities and demonstrate how to apply the KPCA technique into the analysis of earth science data sets. In such applications, a large number of principal components should be retained for studying the spatial patterns, while the variance cannot be quantitatively transferred from the feature space back into the input space. Therefore, we propose a KPCA pattern selection algorithm based on correlations with a given geophysical phenomenon. We demonstrate the algorithm with two widely used data sets in geophysical communities, namely the Normalized Difference Vegetation Index (NDVI) and the Southern Oscillation Index (SOI). The results indicate the new KPCA algorithm can reveal more significant details in spatial patterns than standard PCA.

A RETRIEVAL METHOD OF SIMILAR QUESTION ARTICLES FROM WEB BULLETIN BOARD

Yohei Sakurai, Soichiro Miyazaki and Masanori Akiyoshi

This paper proposes a retrieval method of similar question articles from Web bulletin board, which basically uses cosine similarity index derived from a user's query sentence and article question sentences. Since these sentences are mostly short, it is difficult to distinguish whether article question sentences are similar to a user's query sentence or not simply by applying the conventional cosine similarity index. Therefore our method modifies the elements of word vector used in cosine similarity index, which are derived from a sentence structure from the viewpoints of common words and non-common words between a user's query sentence and article question sentences. Our proposed method is considered to be effective through experiments.

AN EXTRACTION METHOD OF TIME-SERIES NUMERICAL DATA FROM ENTERPRISE PRESS RELEASES

Masanori Akiyoshi, Mayu Gen, Masaki Samejima and Norihisa Komoda

This paper addresses an extraction method of time-series numerical data from enterprise press releases for business strategy design. Business strategy consists of logical actions for continuously producing enterprise outcome. The business strategy design process that is partially based on competitive environment analysis may extremely resort to professional skills so far. To enhance and accelerate the competitive environment analysis, we focus on press releases of competitors in order to extract numerical data related to products or services. Sentences in press releases are well organized and grammatically correct. Therefore such extraction is simply done by identifying the keywords of products or services and the unit indicator co-occurrence. In addition to such simple approach, we clarify the specific rules to applying our method to practical press releases.

PARTNER ASSESSMENT USING MADM AND ONTOLOGY FOR TELECOM OPERATORS

Long Zhang and Xiaoyan Chen

Nowadays, the revenue of telecom operators generated by traditional services declined dramatically while the value added services involving third party value added service providers (partners) are becoming the most prominent source of revenue growth. To regulate the behaviours of the partners and make the operators be able to select best service for end users among the services from different providers, a flexible partner assessment framework is required. This paper 1) presents a flexible partner assessment framework based on Multiple Attribute Decision Making (MADM) method for telecom operators to adapt to the changing requirements of value-added services; 2) proposes ontology to model the complicated relationship in the assessment factors to achieve high extensibility for the continually increasing decision knowledge for partner assessment. From our study, the method adopted and the system proposed can handle the partner assessment problem and support service selection reasonably in telecom industry.

GEOSPATIAL PUBLISHING - Creating and Managing Geo-Tagged Knowledge Repositories

Arno Scharl

International media have recognized the potential of geo-browsers such as NASA World Wind and Google Earth, for example when Web and television coverage on hurricane “Katrina” used interactive geospatial projections to illustrate its path and the scale of destruction. Yet these early applications only hint at the true potential of geo-browsing technology to build and maintain virtual communities, and to revolutionize the production, distribution and consumption of media products. Investigating this potential, this paper discusses geospatial publishing with a special focus on extracting geospatial context from unstructured textual resources. A content analysis of online coverage based on a suite of text mining tools sheds light on the popularity and adoption of geo-browsing platforms. While such platforms might help enrich a company’s portfolio of media products, they also pose a threat for existing players through attracting new competitors; e.g., independent providers of geospatial metadata or location-based services.

MODELLING AND MANAGING KNOWLEDGE THROUGH DIALOGUE: A MODEL OF COMMUNICATION-BASED KNOWLEDGE MANAGEMENT

Violaine Prince

In this paper, we describle a model that relies on the following assumption; ontology negotiation and creation is necessary to make knowledge sharing and KM successful through communication. We mostly focus on the modifying process, i.e. dialogue, and we show a dynamic modification of agents knowledge bases could occur through messages exchanges, messages being knowledge chunks to be mapped with agents KB. Dialogue takes account of both success and failure in mapping. We show that the same process helps repair its own anomalies. We describe an architecture for agents knowledge exchange through dialogue, an instantiation of which has been previously presented in ICEIS2005 Proceedings. Last we conclude about the benefits of introducing dialogue features in knowledge management.

A VIEW ON THE WEB ENGINEERING NATURE OF WEB BASED EXPERT SYSTEMS

Ioannis M. Dokas and Alexandre Alapetite

The Web has become the ubiquitous platform for distributing information and computer services. The tough Web competition, the way people and organizations rely on Web applications, and the increasing user requirements for better services have raised their complexity. Expert systems can be accessed via the Web, forming a set of Web applications known as Web based expert systems. This paper supports that the Web engineering and expert systems principals should be combined when developing Web based expert systems. A development process model will be presented that illustrates, in brief, how these principals can be combined. Based on this model, a publicly available Web based expert system called Landfill Operation Management Advisor (LOMA) was developed. In addition, the results of an accessibility evaluation on LOMA – the first ever reported on Web based expert systems – will be presented. Based on this evaluation some thoughts on accessibility guidelines specific to Web based expert systems will be reported.

SPECIFICATION OF DEPENDENCIES FOR IT SERVICE FAULT MANAGEMENT

Andreas Hanemann, David Schmitz, Patricia Marcu and Martin Sailer

The provisioning of IT services is often based on a variety of resources and underlying services. To deal with this complexity the dependencies between these elements have to be well-known. In particular, dependencies are needed for tracking a failure in a higher-level service being offered to customers down to the provisioning infrastructure. Another usage of dependencies is the impact estimation of an assumed or actual resource failure onto the services to allow for decision making about appropriate measures. Starting from a real-world service provisioning scenario a set of requirements is derived in this paper which has to be addressed by the modeling of dependencies within a service configuration management solution. A subsequent analysis of the state-of-the-art shows the contributions and limitations of existing research approaches and industry tools. To cope with the requirements found earlier, a methodology is proposed to model dependencies for given service provisioning scenarios. Afterwards, some examples are provided for the real-world scenario. The proposed dependency modeling is part of a larger solution for an overall service management information repository.

A MODEL MULTI-AGENTS FOR SHARING AND EXCHANGING KNOWLEDGE IN COMMUNITY OF PRACTICES

Kenfack Clauvice

This paper attempt to show, how communities of practices evolve knowledge transfer. We are focusing on the elaboration of a frame to analyse and underlie the logics of the modalities of functioning of communities of practices. We qualify this concept as an abstract regrouping knowledge creation. By adopting a processual perspective, we will try to present the mechanisms of sharing knowledge within a community of practices (CoPs). The communities of practices is a collection of agents (human beings) who have rather strong common points such as, their level of social capacity, their competences, and the cognitive capacities. The development of the exchanges is based on abstract boundaries; the couple knowledge/community implies that exchanges of information takes place through mechanisms of co-operation, negotiation and through a specific communication language of community members. However, the legitimacy of exchanged knowledge is recognized only with interpersonal confidence association that creates for itself progressively interactions. Besides, these exchanges take place only through rules and standards established by the whole of the members. After having pointed out the theoretical bases of the community concept of practices and the sharing knowledge mechanisms, we will present an approach of simulation using the paradigm of the multi-agents systems, sharing knowledge’s within the community of practices.

SOME SPECIFIC HEURISTICS FOR SITUATION CLUSTERING PROBLEMS

Boris Melnikov, Alexey Radionov, Andrey Moseev and Elena Melnikova

The present work is a continuation of several preceding author's works dedicated to a specific multi-heuristic approach to discrete optimization problems. This paper considers those issues of this multi-heuristic approach which relate to the problems of clustering situations. In particular it considers the issues of the author’s approach to the problems and the description of specific heuristics for the problems. We give the description of a particular example from the group of “Hierarchical Clustering Algorithms”, which we use for clustering situations. We also give descriptions of some common methods and algorithms related to such clustering. There are two examples of metrics on situations sets for two different problems; one of the problems is a classical discrete optimization problem and the other one is a game-playing programming problem.

ROA MODULAR LDAP-BASED APPROACH TO INDUSTRIAL SOFTWARE REVISION CONTROL

Cristina De Castro and Paolo Toppan

A software revision control system stores and manages successive, revised versions of applications, so that every design stage can be easily backtracked. In an industrial context, revision control concerns the evolution of software installed on complex systems and plants, where the need for revision is likely to arise from many different and correlated factors. The concept of revision control must thus be merged in this broader framework, in order to represent all the aspects defining a plant installation lifecycle. In this paper, some schemes are discussed for the definition of a software revision information system representing such factors. An LDAP-based architecture is addressed for modelling and storing their evolution.

AN ACQUISITION KNOWLEDGE PROCESS FOR SOFTWARE DEVELOPMENT

Sandro Ronaldo Bezerra Oliveira, Alexandre Marcos Lins de Vasconcelos, Albérico Lima de Pena Júnior and Lúcio Câmara e Silva

Knowledge must be managed efficiently through the capture, maintenance and dissemination of it in an organization. However, knowledge related to business processes execution is distributed in documents, corporative systems and in key-members minds making the access, preservation and distribution of this knowledge to other members more difficult. In this context, systematic knowledge acquisition processes are necessary to acquire and preserve organizational knowledge. This work presents a process to acquire tacit and explicit organization members’ knowledge related to business processes, and the functionalities of a tool developed to support the execution of this process in a software development context. This tool is part of a software process implementation environment, called ImPProS, developed at CIn/UFPE – Center of Informatics/Federal University of Pernambuco.