Crianças e Propagandas no TikTok: identificando publicidade infantil na rede social TikTok

This study addresses the identification of advertising on TikTok and its impact on children’s behavior, highlighting the growing presence of this platform in children’s routines. Considering the subtle and often disguised nature of advertising aimed at children, the work focuses on identifying and analyzing all advertising done by influencers for children on this social network. Using a machine learning-based methodology, we built a database of influencer posts by automatically classifying advertising and non-advertising content aimed at children. We applied algorithms such as Naive Bayes and Support Vector Machine to classify videos that accurately address child advertising on TikTok. The results showed that the SVM model performed best, with high accuracy and F1-Score. The research also proposes future improvements, including context analysis and the expansion of the keyword base, to enhance the detection of child advertising and protect young users of the platform.