Gender-Based Violence in Paraguay: Selected characteristics and predictive models

Abstract

This study aims to identify the characteristics influencing gender-based violence in Paraguay using machine learning algorithms. Utilizing data from the National Survey on the Situation of Women in Paraguay (ENSIMUP), several predictive models, including Random Forest and Logistic Regression, were applied to analyze factors associated with gender-based violence. The methodology involved rigorous data preprocessing and careful feature selection to enhance model accuracy. The results highlighted the importance of factors such as age, income level, and housing conditions in predicting gender-based violence. The Random Forest model demonstrated superior performance by balancing precision and discriminatory capability. This study underscores the usefulness of predictive models in shaping public policies, suggesting that their integration can significantly improve strategies for preventing and responding to gender-based violence. Continued research with advanced models and data balancing techniques is recommended to optimize risk factor identification and strengthen data-driven policies.