Predicting emerging trends: a machine learning approach to topic popularity on social media

In the dynamic realm of social media, various topics emerge daily, with some evolving into widespread trends. This study focuses on predicting whether a topic will gain popularity in the future, treating this challenge as a binary classification problem. Our approach involves extracting traditional features from topic data, analyzing temporal characteristics, and evaluating textual features to comprehensively represent each topic. Using these multidimensional features, we develop and compare multiple machine learning models to predict topic popularity. The proposed methodology achieves significant accuracy, highlighting its potential for identifying emerging trends on social media platforms. This work provides a robust framework for understanding and forecasting topic popularity, with applications in marketing, public opinion analysis, and content recommendation systems.