Graph deep learning hashtag recommender for reels

Recommendation systems are algorithms that attempt to provide consumers with relevant suggestions for items, such as movies, videos, or reels (micro videos) to watch, hashtags for their posts, songs to listen to, or items to purchase. In many businesses, recommender systems are essential because they can create enormous amounts of revenue and make the platform stand out when compared to others. In this project, we focus on providing users with hashtag recommendations for the reels they want to post, based on their preferences for the reels and hashtags, creating a personalized recommender for each user. Users add hashtags to posts which, in their opinion, properly depict the content in the post. Thus, we should consider both users’ preferences for the content of the post and their preferences for the hashtags. Most methods that implement hashtag recommendations model interactions between users and hashtags or hashtags and posts, unlike this scenario, where we design a hashtag recommendation system based on users, reels, and hashtags. The dataset was built using web scraping, which downloaded the reels from TikTok. We extracted video and audio features from the videos and text features from the hashtags used by users to annotate the reel. We built GNNMCL, a personalized hashtag recommender to recommend hashtags to users based on their previous posts, taking into account both their preferences for the content of the post and their understanding of hashtags. Our GNNMCL graph deep learning model performed better than existing approaches by achieving a NDCG score of 0.9156.