Endereço para acessar este CV: http://lattes.cnpq.br/9872220710484620
Última atualização do currículo em 20/06/2018
Possui graduação em Ingenieur d'Etat en Informatique pela Universidade Abderrahmane Mira de Bejaia (2007), mestrado em Redes e Sistemas Distribuídos pela Universidade Abderrahmane Mira de Bejaia (2010) e doutorado em Ciências da Computação pela Universidade de São Paulo (2016). Tem experiência na área de Ciência da Computação, com ênfase em Banco de Dados, atuando principalmente nos seguintes temas: Métodos de acesso métricos, Busca por similaridade, Indexação de bases de dados e Dados faltantes. (Texto informado pelo autor)
Universidade de São Paulo, Instituto de Ciências Matemáticas e de Computação. Av. Trabalhador São-carlense, 400 Centro 13560970 - São Carlos, SP - Brasil Telefone: (016) 33739700 Fax: (016) 33712238 URL da Homepage: www.icmc.usp.br
Doutorado em Ciências da Computação. Universidade de São Paulo, USP, Brasil.
Título: Techniques for indexing large and complex datasets with missing attribute values, Ano de obtenção: 2016. Orientador: Caetano Traina Junior. Coorientador: Agma Juci Machado Traina. Bolsista do(a): Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, CAPES, Brasil. Palavras-chave: Missing Attribute Values; Similarity Search; Metric access methods; Fractal dimension. Grande área: Ciências Exatas e da Terra
2008 - 2010
Mestrado em Redes e Sistemas Distribuídos. Universidade Abderrahmane Mira de Bejaia, UAMB, Argélia.
Título: Contextual Extraction of Web Content,Ano de Obtenção: 2010. Orientador: Prof. Mohamed Quafafou. Grande área: Ciências Exatas e da Terra
2002 - 2007
Graduação em Ingenieur d'Etat en Informatique. Universidade Abderrahmane Mira de Bejaia, UAMB, Argélia.
Título: Study and Development of an Access Control System for Web Services. Orientador: Dr Abdelkamel Tari.
Pós-Doutorado. Instituto de Ciências Matemáticas e de Computação, São Carlos, ICMC, Brasil.
Bolsista do(a): Conselho Nacional de Desenvolvimento Científico e Tecnológico, CNPq, Brasil. Grande área: Ciências Exatas e da Terra
Study and Development of Techniques for Indexing Large Databases of Complex Data with Handling Missing Values
Descrição: The broad area of multimedia applications involves several problems of data storage and retrieval. Missing data is a common issue for data quality and strongly impacts indexing structures. In fact, both spatial and metric access methods experience poor query performance over databases in the presence of missing values.
Most of the approaches used to handle missing attribute values assume that the data are missing at random, and rely on the statistical and regression models to estimate the missing values, in order to reuse the existing indexing structures and treat the data as if it was never missing. Unfortunately, these techniques are prone to estimation errors, which increase at superlinear rate, both with increasing dimensionalities and with increasing amounts of missing values. Therefore, indexing high-dimensional incomplete data is a challenging task since most of the proposed solutions fail to scale with the amounts
of missing data, and reveal query performance degradation with respect to the ideal situation where all the data items are complete, and a comparison with the sequential scan becomes unavoidable.
The goal of this project is to deﬁne a new model to deal with missing values, underpinned
on the development of a new access method able to index a partially speciﬁed multi-dimensional data. The new model consists of making use of the available information in the incomplete dataset and the correlation property of the data to provide a practical and scalable solution for similarity retrieval over incomplete high-dimensional databases, with efﬁciency and effectiveness indicators comparable to those achieved on complete databases.. Situação: Concluído; Natureza: Pesquisa.
2009 - 2010
Contextual Extraction of Web Content
Descrição: Information retrieval has long been oriented toward textual information, but since the advent of the Web, the document has taken a complex form and has provided multiple types of media. Moreover, the information volume explosion within the Web has created an enormous need for very effective ways to meet the user?s information need. This need has been the subject of several recent studies.
In this project, we suggest a technique for Web page representation based on the content
blocks. This technique belongs to the field of information extraction on the Web. The idea is to segment the Web page into a set of blocks based on its visual structure and build the block indexes using the information held in each content block, and then derive the page index, in order to consider the content blocks instead of the entire page during a query process.
After implementing both of the developed methods, our experimental results show that the suggested technique achieves a very high precision ratio, which means that our technique
of Web page representation makes the research very accurate.. Situação: Concluído; Natureza: Pesquisa. Alunos envolvidos: Mestrado acadêmico: (1) . Integrantes: Safia Brinis - Coordenador / Mohamed Quafafou - Integrante.
2006 - 2007
Study and Development of an Access Control System for Web Services
Descrição: The proliferation of distributed environments has created for large companies a need to expose all or part of an application's functionality to other applications on the Internet or on a private network. One of the major problems is how to communicate and integrate various applications that are independently developed and executed on a variety of hardware platforms and software. This problem is known under the name of "interoperability".
One of the main challenges of the Web Service technology is to ensure the interoperability needed to realize distributed applications. However, to achieve this challenge, the Web Services rely on HTTP or HTTPS and can become tools of circumvention of security policies. Indeed, Web Services belong to the family of "Remote Procedure Calls (RPC)" that has long been a subject of security failure and uncontrolled disclosure of information.
Nowadays, access control systems are present everywhere: from simple door's lock to complex mechanisms of security in the bank's coffers, no place is immune. The same applies to computer systems: any action by a user requires for him/her to be correctly identified with the system to prevent unauthorized access. Mainly, at a time where most of computer resources are interconnected, the need to specify the rights of individuals in a simple and precise way, is a major objective in any access control system.
Many models of access control management have been reported by the security comunity, most of which still represent sever security defects, whether at the level of expressing the specific conditions of access control or at the level of complexity to manage the distributed and dynamic environments. Moreover, most of access control mechanisms are designed to control the immediate access to objects regardless of the implicit information flow divulgated after the dissemination of the information.
This project aims at exploring the Web Service technology and the different mechanisms associated to security, and understanding the organization of access control systems to design and develop the core of a new access control system for Web Services, that uses advanced concepts for flow control information. The resulting core can be implemented in various systems requiring enhaced access control system, such as hospitals, banks and airline companies.. Situação: Concluído; Natureza: Pesquisa. Alunos envolvidos: Graduação: (1) . Integrantes: Safia Brinis - Coordenador / Abdelkamel Tari - Integrante.
Participação em eventos, congressos, exposições e feiras
Workshop of Mining, Indexing and Visualizing Big Data in Clinical Decision Support Systems - (MIVisBD).Similarity Search for Databases with Missing values.
III Workshop dos Pós- Doutorandos do ICMC/USP ? 2017.Similarity Search for Databases with Missing Data.
IEEE International Symposium on Computer-Based Medical Systems. 2015. (Congresso).
Simpósio Brasileiro de Banco de Dados (SBBD).Analyzing Missing Data in Metric Spaces.
Workshop de teses e dissertações (WTD) do Instituto de Ciências Matemáticas e de Computação (ICMC).Estudo e desenvolvimento de técnicas de indexação de grandes bases de dados com valores de atributos faltantes.