Modelo y algoritmo para la identificación y la clasificación de afectaciones en un corpus de testimonios sobre desaparición forzada

This article presents the methodology for the construction of a model and an algorithm for the identification and classification of affectations (economic, physical, political, psychological and sociocultural) in a corpus of testimonies of relatives of victims of enforced disappearance. Initially, the context and characteristics of the phenomenon of enforced disappearance in Colombia are described and the research questions guiding the processing and modeling of the information contained in relation to the volume of data, the conceptual models for the classification and the metrics to measure the efficiency of the implemented algorithm are stated. The methodology and the conceptual basis on which the model is based are then detailed, followed by a discussion of the metrics and their evaluation criteria to compare the efficiency between automatic and manual classification. Finally, reflections and lines of future work on the use of natural language processing techniques in historical memory texts are offered.