Logo_Negativo (2)

Artificial Intelligence Analysis of Youth Migration Intention Survey in Nicaragua

In this project, an analysis of surveys of young Nicaraguans was carried out using the following methodologies Artificial Intelligencewith the main objective of know and understand their migratory intention by generating profiles of individuals based on their roots and various family, socio-economic, environmental perception and future perspective factors. Thus, knowing the main reasons that favour the migration of young people from Nicaragua, the definition of specific policies, strategies, interventions and actions for each profile will be facilitated that seek to prevent or reduce these migrations.

3 methodologies were followed: the first methodology consisted of the identification of respondent profiles and their statistical characterisation in order to group them on the basis of similar situations or realities:

Mapa Autoorganizado (SOM) de calor con frecuencias relativas sobre intención migratoria realizado para Acción Contra el Hambre

In the second metholodogy, a predictive model of migration intentionality based on decision trees was developed to synthesise and predict in a simple way, in this case, the migration intention of a respondent with limited or reduced knowledge of their answers, where the winning prediction (green colour), the percentage of prediction and the percentage of respondents receiving the node are indicated:

Árbol de decisión creado mediante Machine Learning en el que se incluyen las variables de la encuesta de intención migratoria en Nicaragua en el que se predicen las variables que mejor predicen la intención migratoria

The third methodology is a more complete, complex and advanced version of the second one, consisting of the generation of a Random Forest; an iterative model of decision trees processed using Machine Learning Techniques which makes it possible to predict migration intentionality based on the answers given in the questionnaire, which generates a ranking list of the survey variables based on their importance (explainability) to the modelThe aim of this study is to predict the migration intentionality of young people in Nicaragua:

Tabla que representa el modelo de reducción de variables referido a la importancia que cada variable tiene en el modelo
This methodology also allows the above data to be represented by Self-Organising Maps (SOM). In this case, the variable that has obtained the greatest importance in the model is shown and that best predicts migration intentionality: "Migration Transit Factors: Taking the risk of migrating on a regular basis":
Mapas Autoorganizados SOM correspondiente a la variable que ha obtenido la mayor importancia en el modelo Random Forest sobre predicción de la intencionalidad migratoria con Inteligencia Artificial
The results After carrying out the study, they found 6 exclusive and excluding profiles of young migrants with one or several defining factors for each of them, which provide indicators to help the Nicaraguan administration make decisions regarding the migration of young people in their country.