Algorithm for Safety Decisions in Social Media Feeds Using Personification Patterns

Abstract

For safety decisions in social media applications, it is necessary to classify personification patterns. The paper proposes using video material to apply machine learning to select, and extract significant feature qualities and grasp the semantics of feature space connection to comprehend the personification of a certain user. The feature traits are based on a computer vision-based approach and a natural language-based approach. A strong belief is calculated from language descriptions and persona traits. These traits are then used to determine the overlap of feature space using various ML algorithms to deduce the intrinsic relationships. The proposed goal is validated by this algorithm and user personification is an important aspect that can be captured through video analytics. Using this personification-based method, better decisions can be made in the given domain space.