Text Analysis Software Using Topic Modeling Techniques for the Extraction of Knowledge from Cases Related to Vulnerability and Access to Justice
Cargando...
Compartir
Fecha
Autores
Espinosa J.E
Mateus S.P
Ramirez D.M.
Título de la revista
ISSN de la revista
Título del volumen
Editor
Springer Science and Business Media Deutschland GmbH
Resumen
Descripción
Access to justice is a vital part of the UNDP (United Nations Development Programme) mandate to reduce poverty and strengthen democratic governance. One of the objectives of the Personero (ombudsmen) in emerging countries is to assist people in accessing justice, promoting equality and the protection of the most vulnerable, providing assistance to citizens through guidance or consultations. In Medellín (Colombia), the “Personería de Medellín” (ombudsman’s office) has established a citizen service system staffed by lawyers who address their difficulties, needs, or other issues, guiding them on the steps to take, such as filing legal actions or providing advice. All this information is recorded in a digitalized system. Annually, more than 16,000 different cases can be recorded (e.g. for year 2019), and despite the implementation of multidimensional analysis tools like dashboards, there is a need to implement Natural Language Processing (NLP) techniques to semantically analyze the fields of unstructured text information. This paper presents an application offered to citizens called Person_IA, which implements various functionalities based on artificial intelligence techniques, particularly Natural Language Processing (NLP). This allows for text analysis using strategies such as topic modeling, word clouds, n-grams, among others, to extract knowledge from the unstructured information in the citizen service system. This helps generate indicator reports and analyze variables that enable efficient decision-making by knowledge managers at the Medellín Ombudsman’s Office, facilitating access to justice for the vulnerable population in this city. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
Palabras clave
Latent Dirichlet Assignment (LDA), Latent Semantic Analysis (LSA), Machine Learning, Natural Language Processing (NLP), Textual Analysis, Topic Modelling, Behavioral research, Data mining, Decision making, Laws and legislation, Learning algorithms, Modeling languages, Natural language processing systems, Semantics, Dirichlet, Language processing, Latent dirichlet assignment, Latent semantic analyse, Latent Semantic Analysis, Machine-learning, Natural language processing, Natural languages, Textual-analysis, Topic Modeling, Machine learning