1. Inicio
  2. Atrás
Título: E-mail processing with fuzzy SOMs and association rules
Autores: Lanzarini, Laura Cristina
Villa Monte, Augusto
Estrebou, César
Fecha: 2011-03-31
2011
Publicador: Unversidad Nacional de La Plata
Fuente:


Tipo: Articulo
Articulo
Tema: information retrieval; data mining; text mining; e-mails analysis; FSOM; association rules
Ciencias Informáticas
Almacenamiento y Recuperación de la Información
Procesamiento Automatizado de Datos
Internet
Descripción: E-mail texts are hard to process due to their short length. In this article, the use of a diffuse neural network that is capable of identifying the most relevant terms in a set of e-mails is proposed. The associations between these terms will be measured through association rules built with the terms identified by the network. The metrics support, confidence and interest of the rules will be used to qualify the corresponding terms. The method proposed has been used to process e-mails of the PACENI Project (Support Project for Improving First-Year Teaching in Courses of Studies in Exact and Natural Sciences, Economic Science and Computer Science). With this type of analysis, the most common topics of student questions have been identified. Even though this new information can have various applications, they all involve, as a first instance, an improvement in student service.
Idioma: Inglés