Topic modeling for social media content: A practical approach

Perceiving the discussed topic in social media brings a great amount of value to different fields, such as marketing, security, education, and management. Topic modeling provides a powerful method for projecting text documents into topic space. In this paper, we explore an unsupervised topic modeling approach that incorporates LDA algorithm toward discovering the topics in social media content. Empirical experiments on social media datasets with 90,527 records reveal that this approach is quite effective for detecting the topic facets and extracting their dynamics over time. By analyzing the studied dataset, five main topics were discovered accurately by the presented algorithm according to the domain experts' comments. The presented model is quite general and can be applied in a wide variety of domains to automatically mining topics from the social media channels.