Estella Annoni - Academia.edu (original) (raw)

Papers by Estella Annoni

Distributed software environments are increasingly complex and difficult to manage, as they integ... more Distributed software environments are increasingly complex and difficult to manage, as they integrate various legacy software with specific management interfaces. Moreover, the fact that management tasks are performed by humans leads to many configuration errors and low reactivity. This is particularly true in medium or large-scale distributed infrastructures. To address this issue, we explore the design and implementation of an autonomic management system. The main principle is to wrap legacy software pieces in components in order to administrate a software infrastructure as a component architecture. However, we observed that the interfaces of a component model are too low-level and difficult to use. Consequently, we explore the use of a model driven approach where several UML profiles are used to specify the different facets of an autonomic management policy.

The development of software facilitating decision taking is becoming more frequent, due to the gr... more The development of software facilitating decision taking is becoming more frequent, due to the growing need of reactivity and competitiveness within companies. Such software is called Decision Support Systems (DSS). However, 80% of decision making projects fail to satisfy user requirements, and 40% fail to help decision making 3. The recent works of methods bound to DSS define their schemes from the user requirements and the source systems. However, they can not represent all specificities of the DSS. Moreover, they rely on specific models representing only data, the dynamic aspect of DSS being largely ignored. Thus, none of these models is accepted neither by the researchers nor the practitioners.

Traditionally, mining web page contents involves modeling their contents to discover the underlyi... more Traditionally, mining web page contents involves modeling their contents to discover the underlying knowledge. Data extraction proposals represent web data in a formal structure such as database structures specific to application domains. Those models fail to catch the full diversity of web data structures which can be composed of different types of contents, and can be also unstructured. In fact, with these proposals, it is not possible to focus on a given type of contents, to work on data of different structures and to mine on data of different application domains as required to mine efficiently a given content type or web documents from different domains. On top of that, since web pages are designed to be understood by users, this paper considers modeling of web document presentations expressed through HTML tag attributes as useful for an efficient web content mining. Hence, this paper provides a general framework composed of an object-oriented web data model based on HTML tags and algorithms for web content and web presentation object extraction from any given web document. From the HTML code of a web document, web objects are extracted for mining, regardless of the domain.

Proceedings of the 11th International Conference on Enterprise Information

Traditionally, mining web page contents involves modeling their contents to discover the underlyi... more Traditionally, mining web page contents involves modeling their contents to discover the underlying knowledge. Data extraction proposals represent web data in a formal structure such as database structures specific to application domains. Those models fail to catch the full diversity of web data structures which can be composed of different types of contents, and can be also unstructured. In fact, with these proposals, it is not possible to focus on a given type of contents, to work on data of different structures and to mine on data of different application domains as required to mine efficiently a given content type or web documents from different domains. On top of that, since web pages are designed to be understood by users, this paper considers modeling of web document presentations expressed through HTML tag attributes as useful for an efficient web content mining. Hence, this paper provides a general framework composed of an object-oriented web data model based on HTML tags and algorithms for web content and web presentation object extraction from any given web document. From the HTML code of a web document, web objects are extracted for mining, regardless of the domain.

The development of software facilitating decision taking is becoming more frequent, due to the gr... more The development of software facilitating decision taking is becoming more frequent, due to the growing need of reactivity and competitiveness within companies. Such software is called Decision Support Systems (DSS). However, 80% of decision making projects fail to satisfy user requirements, and 40% fail to help decision making 3. The recent works of methods bound to DSS define their schemes from the user requirements and the source systems. However, they can not represent all specificities of the DSS. Moreover, they rely on specific models representing only data, the dynamic aspect of DSS being largely ignored. Thus, none of these models is accepted neither by the researchers nor the practitioners.


, http://www.irit.fr Résumé. Les systèmes d'information décisionnels (SID) sont des systèmes d'in... more , http://www.irit.fr Résumé. Les systèmes d'information décisionnels (SID) sont des systèmes d'information (SI) qui ont pour objectif de faciliter la prise de décision à partir d'information résultant de processus complexes de dérivation et de préparation des données de SI sources. Ces processus sont généralement peu modélisés et sont directement implantés avec des logiciels spécifiques au cours des projets décisionnels bien que trois modèles particuliers ont été proposés pour représenter ces processus. En effet, ces modèles utilisent de nouvelles notations distinctes de celles de la modélisation des données qu'ils proposent. Ils requièrent deux schémas distincts pour les données et les processus alors que les schémas conceptuels des SID sont déjà nombreux et énormes en raison de la taille des projets et des spécificités des domaines. Ainsi, nous proposons des outils pour la prise en compte de la dynamique des SID et une modélisation intégrée des processus dans le schéma des SID. Nous définissons les spécificités de la dynamique des SID. L'objectif est de prendre en compte dans un unique schéma aussi bien les processus liés à la dérivation des données étudiés par les travaux précédents que ceux liés à la préparation des données de l'environnement de prise de décision.


La démocratisation des systèmes d'information décisionnels (SID) nécessite le développement de mé... more La démocratisation des systèmes d'information décisionnels (SID) nécessite le développement de méthodes de conception. Contrairement aux modèles de systèmes d'information (SI) qui n'ont pas pour objet d'être compris par les utilisateurs, les modèles des SID doivent être exploitables par les analystes et les décideurs. Parmi les méthodes d'ingénierie des SID qui ont été proposées, rares sont celles qui explicitent la tâche d'analyse des besoins. Pour ces raisons, nous proposons une démarche de collecte et de formalisation des besoins des utilisateurs du SID en utilisant des modèles proches de leur vision des données. A partir des besoins spécifiés sous forme de tableaux multidimensionnels, nous proposons des extensions de la modélisation objet afin de formaliser les besoins en terme de données et de traitements dans le contexte multidimensionnel.

Fifth International Conference on Information Technology: New Generations (itng 2008), 2008

Distributed software environments are increasingly complex and difficult to manage, as they integ... more Distributed software environments are increasingly complex and difficult to manage, as they integrate various legacy software with specific management interfaces. Moreover, the fact that management tasks are performed by humans leads to many configuration errors and low reactivity. This is particularly true in medium or large-scale distributed infrastructures. To address this issue, we explore the design and implementation of an autonomic management system. The main principle is to wrap legacy software pieces in components in order to administrate a software infrastructure as a component architecture. However, we observed that the interfaces of a component model are too low-level and difficult to use. Consequently, we explore the use of a model driven approach where several UML profiles are used to specify the different facets of an autonomic management policy.

Lecture Notes in Computer Science, 2006

ABSTRACT

Ingénierie des systèmes d'information, 2005

Decision designers express more and more the need of methods because of the important use of Deci... more Decision designers express more and more the need of methods because of the important use of Decision Support Systems (DSS). The lack of norms implies development of several concepts and methods in decision-making domain. However, there are some standards therefore it is necessary to reuse the existent. In this paper, we describe two main components of the method we define: a pattern system and an approach of DSS design. Our method allows quick and reliable development of DSS complete compared to user requirements. It focuses on reuse and it creates DSS according to user requirements while integrating various architectures. Moreover, it evaluates operational systems in relation to user requirements before the step of design.

Data Warehousing and Knowledge Discovery, 2006

Towards Multidimensional Requirement Design Estella Annoni, Franck Ravat, Olivier Teste, and Gill... more Towards Multidimensional Requirement Design Estella Annoni, Franck Ravat, Olivier Teste, and Gilles Zurfluh ... Luján-Mora, S., Vassiliadis, P., Trujillo, J.: Data mapping diagrams for data warehouse de-sign with uml. In Atzeni, P., Chu, WW, Lu, H., Zhou, S., Ling, TW, eds.: ER. ...

Les systèmes d'information décisionnels (SID) ont pour objectif de faciliter la prise de décision... more Les systèmes d'information décisionnels (SID) ont pour objectif de faciliter la prise de décision. Ils reposent principalement sur la modélisation multidimensionnelle où le sujet de l'analyse, appelé fait, avec ses mesures sont représentés au centre d'une étoile dont les branches sont les dimensions de l'analyse avec ses paramètres. De nombreux modèles ont été proposés, mais ils permettent de représenter uniquement les liens entre les dimensions et les paramètres. Certains liens entre les faits ont récemment été étudiés. Les liens entre mesures représentent les corrélations propres à une activité car elles sont les caractéristiques de celle-ci. Mais, il n'existe pas de travaux définissant et modélisant ces liens. En ce sens, nous proposons une modélisation des faits et des mesures qui représente les liens entre ces concepts. Notre modélisation utilise UML car il permet de définir une représentation associée et adaptée à la terminologie des SID et de s'adresser au plus grand nombre de concepteurs décisionnels de part sa notoriété. ABSTRACT. Decision support system (DSS) goal is to help decision-makers. They are modeled mainly in a multdimiensional way, that means activity analysis subjects, called facts, with their measures are represented as the center of a star with dimensions and their parameters around. Several models have been devoted to DSS representation, but they handle only relationships between dimensions and their parameters. Recently, some relationships between facts have been analyzed. Relationships between measures indicate correlations through a given activity because they represent properties of this activity. However, there is no work on these relationships. Therefore, we model facts and measures such as relationships between these concepts. We define a UML-based model because it allows a representation related and adapted to DSS terminology and it is well-known by a large amount of designers.

La société I-D6 spécialisée dans le domaine de l'ingénierie des systèmes d'information décisionne... more La société I-D6 spécialisée dans le domaine de l'ingénierie des systèmes d'information décisionnels (SID) a exprimé le besoin primordial d'une méthode de développement de ce type de systèmes. Cette méthode doit permettre un développement rapide de SID fiables et complets par rapport aux besoins des utilisateurs dans un contexte orienté réutilisation. Parmi les méthodes d'ingénierie spécifiques aux SID existantes, aucune n'est communément admise. De plus, elles ne favorisent pas la réutilisation des connaissances.

Distributed software environments are increasingly complex and difficult to manage, as they integ... more Distributed software environments are increasingly complex and difficult to manage, as they integrate various legacy software with specific management interfaces. Moreover, the fact that management tasks are performed by humans leads to many configuration errors and low reactivity. This is particularly true in medium or large-scale distributed infrastructures. To address this issue, we explore the design and implementation of an autonomic management system. The main principle is to wrap legacy software pieces in components in order to administrate a software infrastructure as a component architecture. However, we observed that the interfaces of a component model are too low-level and difficult to use. Consequently, we explore the use of a model driven approach where several UML profiles are used to specify the different facets of an autonomic management policy.

The development of software facilitating decision taking is becoming more frequent, due to the gr... more The development of software facilitating decision taking is becoming more frequent, due to the growing need of reactivity and competitiveness within companies. Such software is called Decision Support Systems (DSS). However, 80% of decision making projects fail to satisfy user requirements, and 40% fail to help decision making 3. The recent works of methods bound to DSS define their schemes from the user requirements and the source systems. However, they can not represent all specificities of the DSS. Moreover, they rely on specific models representing only data, the dynamic aspect of DSS being largely ignored. Thus, none of these models is accepted neither by the researchers nor the practitioners.

Traditionally, mining web page contents involves modeling their contents to discover the underlyi... more Traditionally, mining web page contents involves modeling their contents to discover the underlying knowledge. Data extraction proposals represent web data in a formal structure such as database structures specific to application domains. Those models fail to catch the full diversity of web data structures which can be composed of different types of contents, and can be also unstructured. In fact, with these proposals, it is not possible to focus on a given type of contents, to work on data of different structures and to mine on data of different application domains as required to mine efficiently a given content type or web documents from different domains. On top of that, since web pages are designed to be understood by users, this paper considers modeling of web document presentations expressed through HTML tag attributes as useful for an efficient web content mining. Hence, this paper provides a general framework composed of an object-oriented web data model based on HTML tags and algorithms for web content and web presentation object extraction from any given web document. From the HTML code of a web document, web objects are extracted for mining, regardless of the domain.

Proceedings of the 11th International Conference on Enterprise Information

Traditionally, mining web page contents involves modeling their contents to discover the underlyi... more Traditionally, mining web page contents involves modeling their contents to discover the underlying knowledge. Data extraction proposals represent web data in a formal structure such as database structures specific to application domains. Those models fail to catch the full diversity of web data structures which can be composed of different types of contents, and can be also unstructured. In fact, with these proposals, it is not possible to focus on a given type of contents, to work on data of different structures and to mine on data of different application domains as required to mine efficiently a given content type or web documents from different domains. On top of that, since web pages are designed to be understood by users, this paper considers modeling of web document presentations expressed through HTML tag attributes as useful for an efficient web content mining. Hence, this paper provides a general framework composed of an object-oriented web data model based on HTML tags and algorithms for web content and web presentation object extraction from any given web document. From the HTML code of a web document, web objects are extracted for mining, regardless of the domain.

The development of software facilitating decision taking is becoming more frequent, due to the gr... more The development of software facilitating decision taking is becoming more frequent, due to the growing need of reactivity and competitiveness within companies. Such software is called Decision Support Systems (DSS). However, 80% of decision making projects fail to satisfy user requirements, and 40% fail to help decision making 3. The recent works of methods bound to DSS define their schemes from the user requirements and the source systems. However, they can not represent all specificities of the DSS. Moreover, they rely on specific models representing only data, the dynamic aspect of DSS being largely ignored. Thus, none of these models is accepted neither by the researchers nor the practitioners.


, http://www.irit.fr Résumé. Les systèmes d'information décisionnels (SID) sont des systèmes d'in... more , http://www.irit.fr Résumé. Les systèmes d'information décisionnels (SID) sont des systèmes d'information (SI) qui ont pour objectif de faciliter la prise de décision à partir d'information résultant de processus complexes de dérivation et de préparation des données de SI sources. Ces processus sont généralement peu modélisés et sont directement implantés avec des logiciels spécifiques au cours des projets décisionnels bien que trois modèles particuliers ont été proposés pour représenter ces processus. En effet, ces modèles utilisent de nouvelles notations distinctes de celles de la modélisation des données qu'ils proposent. Ils requièrent deux schémas distincts pour les données et les processus alors que les schémas conceptuels des SID sont déjà nombreux et énormes en raison de la taille des projets et des spécificités des domaines. Ainsi, nous proposons des outils pour la prise en compte de la dynamique des SID et une modélisation intégrée des processus dans le schéma des SID. Nous définissons les spécificités de la dynamique des SID. L'objectif est de prendre en compte dans un unique schéma aussi bien les processus liés à la dérivation des données étudiés par les travaux précédents que ceux liés à la préparation des données de l'environnement de prise de décision.


La démocratisation des systèmes d'information décisionnels (SID) nécessite le développement de mé... more La démocratisation des systèmes d'information décisionnels (SID) nécessite le développement de méthodes de conception. Contrairement aux modèles de systèmes d'information (SI) qui n'ont pas pour objet d'être compris par les utilisateurs, les modèles des SID doivent être exploitables par les analystes et les décideurs. Parmi les méthodes d'ingénierie des SID qui ont été proposées, rares sont celles qui explicitent la tâche d'analyse des besoins. Pour ces raisons, nous proposons une démarche de collecte et de formalisation des besoins des utilisateurs du SID en utilisant des modèles proches de leur vision des données. A partir des besoins spécifiés sous forme de tableaux multidimensionnels, nous proposons des extensions de la modélisation objet afin de formaliser les besoins en terme de données et de traitements dans le contexte multidimensionnel.

Fifth International Conference on Information Technology: New Generations (itng 2008), 2008

Distributed software environments are increasingly complex and difficult to manage, as they integ... more Distributed software environments are increasingly complex and difficult to manage, as they integrate various legacy software with specific management interfaces. Moreover, the fact that management tasks are performed by humans leads to many configuration errors and low reactivity. This is particularly true in medium or large-scale distributed infrastructures. To address this issue, we explore the design and implementation of an autonomic management system. The main principle is to wrap legacy software pieces in components in order to administrate a software infrastructure as a component architecture. However, we observed that the interfaces of a component model are too low-level and difficult to use. Consequently, we explore the use of a model driven approach where several UML profiles are used to specify the different facets of an autonomic management policy.

Lecture Notes in Computer Science, 2006

ABSTRACT

Ingénierie des systèmes d'information, 2005

Decision designers express more and more the need of methods because of the important use of Deci... more Decision designers express more and more the need of methods because of the important use of Decision Support Systems (DSS). The lack of norms implies development of several concepts and methods in decision-making domain. However, there are some standards therefore it is necessary to reuse the existent. In this paper, we describe two main components of the method we define: a pattern system and an approach of DSS design. Our method allows quick and reliable development of DSS complete compared to user requirements. It focuses on reuse and it creates DSS according to user requirements while integrating various architectures. Moreover, it evaluates operational systems in relation to user requirements before the step of design.

Data Warehousing and Knowledge Discovery, 2006

Towards Multidimensional Requirement Design Estella Annoni, Franck Ravat, Olivier Teste, and Gill... more Towards Multidimensional Requirement Design Estella Annoni, Franck Ravat, Olivier Teste, and Gilles Zurfluh ... Luján-Mora, S., Vassiliadis, P., Trujillo, J.: Data mapping diagrams for data warehouse de-sign with uml. In Atzeni, P., Chu, WW, Lu, H., Zhou, S., Ling, TW, eds.: ER. ...

Les systèmes d'information décisionnels (SID) ont pour objectif de faciliter la prise de décision... more Les systèmes d'information décisionnels (SID) ont pour objectif de faciliter la prise de décision. Ils reposent principalement sur la modélisation multidimensionnelle où le sujet de l'analyse, appelé fait, avec ses mesures sont représentés au centre d'une étoile dont les branches sont les dimensions de l'analyse avec ses paramètres. De nombreux modèles ont été proposés, mais ils permettent de représenter uniquement les liens entre les dimensions et les paramètres. Certains liens entre les faits ont récemment été étudiés. Les liens entre mesures représentent les corrélations propres à une activité car elles sont les caractéristiques de celle-ci. Mais, il n'existe pas de travaux définissant et modélisant ces liens. En ce sens, nous proposons une modélisation des faits et des mesures qui représente les liens entre ces concepts. Notre modélisation utilise UML car il permet de définir une représentation associée et adaptée à la terminologie des SID et de s'adresser au plus grand nombre de concepteurs décisionnels de part sa notoriété. ABSTRACT. Decision support system (DSS) goal is to help decision-makers. They are modeled mainly in a multdimiensional way, that means activity analysis subjects, called facts, with their measures are represented as the center of a star with dimensions and their parameters around. Several models have been devoted to DSS representation, but they handle only relationships between dimensions and their parameters. Recently, some relationships between facts have been analyzed. Relationships between measures indicate correlations through a given activity because they represent properties of this activity. However, there is no work on these relationships. Therefore, we model facts and measures such as relationships between these concepts. We define a UML-based model because it allows a representation related and adapted to DSS terminology and it is well-known by a large amount of designers.

La société I-D6 spécialisée dans le domaine de l'ingénierie des systèmes d'information décisionne... more La société I-D6 spécialisée dans le domaine de l'ingénierie des systèmes d'information décisionnels (SID) a exprimé le besoin primordial d'une méthode de développement de ce type de systèmes. Cette méthode doit permettre un développement rapide de SID fiables et complets par rapport aux besoins des utilisateurs dans un contexte orienté réutilisation. Parmi les méthodes d'ingénierie spécifiques aux SID existantes, aucune n'est communément admise. De plus, elles ne favorisent pas la réutilisation des connaissances.