Filtering of Brand – Related Microblogs Using Social – Smooth Multi – View Embedding

Abstract

Filtering of Brand-related Microblogs THE popularity of microblogging platforms such as Twitter1 and Sina Weibo encouraged users to generate and share huge amounts of user-generated social media content (UGCs).Projects on Brand-related Microblog Filtering These UGCs offer real-time information resources for a wide range of topics and benefit a wide range of users and applications.

As a result, extensive research has focused on analyzing social media, such as summarizing social events, analyzing social networks, social television, and sensing topics.

Among social content, brand-related microblogs, which spread much faster than traditional media content, have significant marketing values for businesses and government organizations.

Conclusion

We have proposed a microblog filtering method which can be used for the collection of brand data as a noise filtering step.

Projects on Filtering of Brand-related Microblogs The key component of our method is a discriminatory social-aware embedding approach that maps the content of microblogs, consisting of three (or more) views, into a latent space while taking into account the brand information and social relations of microblogs.

Extensive experiments conducted in the BSN dataset with 100 famous brands have shown that the proposed microblog filtering method can achieve better performance in comparison with state-of – the-art methods.

We also discovered from the experiments an interesting property of social information that it makes more impact on microblog filtering for brands that have influential users with large social connections and follow-ups.