Monday 03 March 2025
A new approach has been proposed in the field of artificial intelligence, which aims to improve the accuracy of user influence level prediction on social media platforms. This is achieved by incorporating insights from socio-linguistic research, behavioral sciences, and computational models.
The study focused on Reddit, a popular online community where users share their thoughts and opinions on various topics. The researchers used a combination of text classification algorithms and user- centric information to predict the influence level of each user. This was done by training separate models for different subreddits, which are communities within Reddit that focus on specific topics.
The results showed that the multi-task model, which learned to perform multiple tasks simultaneously, outperformed the baseline models in predicting user influence levels. The model achieved a mean RankDCG score of 24.85%, which is significantly higher than the scores obtained by the baseline models.
One of the key findings of the study was that the user’s comment alone can be used to predict their influence level with a high degree of accuracy. This suggests that online behavior, such as the content and tone of one’s comments, can be a strong indicator of a user’s influence level.
The researchers also found that different subreddits have distinct characteristics that affect the prediction of user influence levels. For example, users who are active in the AskMen subreddit tend to have higher influence levels than those who are active in other subreddits.
Another interesting observation was that the loss weights for each sub-task were domain-specific. This means that the importance of different factors, such as age and gender, varied across different subreddits.
The study’s findings have implications for social media platforms, which can use this information to personalize content recommendations and improve user engagement. For example, a platform could recommend content to users based on their influence level and the topics they are interested in.
The researchers also highlighted some limitations of their approach, including the reliance on publicly available data and the potential biases present in the dataset. They noted that future studies should aim to address these limitations and explore new approaches to user influence level prediction.
Overall, this study demonstrates the potential of combining insights from socio-linguistic research, behavioral sciences, and computational models to improve user influence level prediction on social media platforms. The findings have significant implications for social media platforms and highlight the importance of understanding online behavior in order to personalize content recommendations and improve user engagement.
Cite this article: “Predicting User Influence on Social Media Platforms”, The Science Archive, 2025.
Artificial Intelligence, Social Media, User Influence Level, Reddit, Text Classification, Multi-Task Model, Rankdcg Score, Online Behavior, Socio-Linguistic Research, Behavioral Sciences
