Modelado basado en datos para la clasificación semiautomática de correspondencia electrónica: caso de estudio para la Administración Pública Colombiana

Vargas Antolínez, Edwin Alberto | 2018

E-mail communication is still the most prevalent form of Customer support process in Organizations. Therefore, organizations have had to implementing processes focused on email categorization in accordance with email contents, in order to provide an efficient response to customer requests. A common text mining approach involves a representation of text based on keywords later combined with machine learning. This project presents a methodological approach that evaluates classification algorithms: Supoport Vector Machine and Gradient Boodting Trees on a corpus builded from emails database of the Administrative Department of Public Function in Colombia.