Cyberbullying prevention model and design for individuals

In this thesis, a cyberbullying prevention model and design has been realized to detect and prevent cyberbullying, which has emerged with the development and spread of technology and causes negative effects on the psychological and mental health of individuals. In the developed model, the audio data obtained by listening to the environment using speech recognition systems were converted into text data. The text data obtained were classified using the Google BERT algorithm. Cyberbullying is a type of bullying that is carried out through digital technologies such as social media, messaging platforms, gaming platforms and mobile phones. It is carried out through behaviors aimed at intimidating, angering, embarrassing and humiliating the bullied people. Cyberbullying has negative psychological and physical effects on individuals and it is seen that they fail in academic and social areas. Considering the effects of cyberbullying in terms of mental, psychological, family-friend relationships and academic success, it cannot be ignored how much its detection and prevention affect the negative consequences that may occur. Cyberbullying can usually be done verbally or in writing on digital platforms. Since the information that comes in the form of voice comes as data that can be converted into voice, it needs additional processes to reveal it. In this context, Google Speech API is generally used to convert voice data into written text, while Google Bert Algorithm can be used to make sense of written texts. Detecting and preventing cyberbullying is of great importance. With this study, which aims to minimize the level and duration of exposure to cyberbullying, it is planned to make it possible to create a security for individuals.