Cybervictim vs. cyberaggressor. profile determination and comparison through artificial neural networks

Abstract

This research aims to determine the magnitude to which aspects of the sociodemographic profile of students (gender, age, religion, ethnicity and race) influence both the acts of harassment received as a cybervictim and those carried out as a cyberaggressor. Additionally, aiming to discover which of these acts suffered as cybervictim or carried out as cyberaggressor increase or decrease to a greater extent, comparing a profile of potential cyberaggressor with another of potential cybervictim. For this purpose, an artificial neural network of the multilayer perceptron type is employed, generating an estimation model of these cybervictim and cyberaggressor facts based on the students’ profile using data obtained through a five-points’ Likert questionnaire. That is especially useful for detecting specific behaviours based on a certain profile, being able to determine which of these attitudes are most at risk of occurring to put the relevant measures in place to prevent them.