Leandro M. Almeida

Home | ContactCurriculum VitaePublications | Ph.D. ProjectM.Phil. ProjectB.Sc. Projects

B.Sc. Projects

Final Course Work (Trabalho de Graduação)

Year 2004
Title: Text Message Categorization for Collaborative Learning Environments. (Categorização de Mensagens de Texto para Ambientes de Aprendizado Colaborativo).
SupervisorDr. Thereza Patrícia P. Padilha.
Keywords: Data Mining, Text Mnining, Chat Messages, Collaborative Learning Environments.

Automatic text categorization is a task in the field of Text Mining that has received a considerable attention from many researchers because the volume of digital information increases on a daily basis. This task essentially categorizes documents or messages with one or more predefined categories. Thus, users can identify the topic of documents without reading them completely and, consequently, prioritize the reading of more relevant documents.

In collaborative learning environments, there is a large amount of text messages exchanged between students during task execution and, unfortunately, professors do not know their real intentions. Student intention identification has vast importance in measuring learning performance. This work presents the project and implementation of a system for student intention identification through message categorization using vector similarity, specifically Euclidean distances. Experimental results for each system module were determined using a simple set of messages from the RESOLVE collaborative learning environment database.

Undergraduate Research Training (Iniciação Científica)

CEULP/ULRBA Scholarship - 2002 to 2004.
Supervisor: Dr. Thereza Patrícia P. Padilha.
Funding: Initiation Program in Science and Technology of the Lutheran University Center of Palmas.

From 2002 to 2003
Title: An Intelligent Agent Learning Model based on Data Mining (Um Modelo de Aprendizado de Agentes Inteligentes Baseado em Técnicas de Data Mining).
Keywords: Data Mining, Intelligent Agents, C4.5.
Abstract: An intelligent agent learning model was developed using data mining methods applied to discover the collaboration profile between users that employ a chat tool in an education setting.

From 2003 to 2004

Title: Development of Intelligent Agents for the RESOLVE Collaborative Learning Environment: agents with collaborative profile (Desenvolvimento de Agentes Inteligentes para o Ambiente de Aprendizado Colaborativo RESOLVE: Agente de Perfil Colaborativo)
Keywords: Intelligent Agents, Data Mining, Collaborative Learning Environment.
Abstract: The purpose of the project was to implant intelligent agents in the RESOLVE collaborative learning environment to assist educators in monitoring the activities of students as well as to facilitate the interaction of groups and provide a more social environment for distance learning.

Title: Aspect Mining: concepts, methods and tools (Mineração de Aspectos: conceitos, métodos e ferramentas)
Keywords: Aspect Mining, Aspect-oriented programming, AspectJ.
Abstract: Object-oriented programming in the scope of system development does not provide a complete treatment for complex systems. This treatment reveals functionalities (aspects) that affect the behavior of a system throughout its entire extension. The field of Aspect Mining has emerged from the need to identify desirable aspects in complex systems, mainly legacy systems. This paper presents an overview of Aspect Mining, focusing on related concepts of aspect-oriented programming, describing existing mining methods and describing the functioning of the AMT/AMTEX and FEAT mining tools. Experiments were carried out with these two tools using JHotDraw framework in order to investigate candidate codes for aspects.

Locations of visitors to this page