who are unprotected.
We promote positive parenting strategies with the double objective of fostering a family bond based on affection and communication, and calling attention to those violent practices naturalized in families.
We also provide children with a safe space to express themselves by means of psychological techniques, such as storytelling and projective drawing.
By predictive analytics and machine learning techniques, we developed a model that identifies households with children at risk of violence. Our preliminary results suggest that up to 40% of at-risk children can be identified, i.e. a 20-fold improvement over current detection levels.
Abrazar uses innovative statistical methodologies that make it possible to prevent and identify cases of child violence before they become serious.
By using machine learning techniques instead of waiting for referrals, local governments will be able to use these tools to find children who are at risk and prevent violence.