Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Forecasting the Buzz: Enriching Hashtag Popularity Prediction with LLM Reasoning

Yifei Xu, Jiaying Wu, Herun Wan, Yang Li, Zhen Hou, Min-Yen Kan

TL;DR

This work tackles hashtag popularity forecasting by integrating reasoning from large language models with traditional regression. The proposed BuzzProphet framework converts LLM-generated rationales about topic category, audience, and timing into enriched features for a lightweight regressor, bridging contextual insight with numeric precision. Evaluation on the HashView Weibo benchmark demonstrates that reasoning-augmented inputs improve accuracy (RMSE, SRC) across multiple regressors, while also yielding human-readable explanations. The approach offers a scalable, interpretable solution for social media trend forecasting and highlights the value of using LLMs as reasoning engines rather than direct numeric predictors.

Abstract

Hashtag trends ignite campaigns, shift public opinion, and steer millions of dollars in advertising spend, yet forecasting which tag goes viral is elusive. Classical regressors digest surface features but ignore context, while large language models (LLMs) excel at contextual reasoning but misestimate numbers. We present BuzzProphet, a reasoning-augmented hashtag popularity prediction framework that (1) instructs an LLM to articulate a hashtag's topical virality, audience reach, and timing advantage; (2) utilizes these popularity-oriented rationales to enrich the input features; and (3) regresses on these inputs. To facilitate evaluation, we release HashView, a 7,532-hashtag benchmark curated from social media. Across diverse regressor-LLM combinations, BuzzProphet reduces RMSE by up to 2.8% and boosts correlation by 30% over baselines, while producing human-readable rationales. Results demonstrate that using LLMs as context reasoners rather than numeric predictors injects domain insight into tabular models, yielding an interpretable and deployable solution for social media trend forecasting.

Forecasting the Buzz: Enriching Hashtag Popularity Prediction with LLM Reasoning

TL;DR

This work tackles hashtag popularity forecasting by integrating reasoning from large language models with traditional regression. The proposed BuzzProphet framework converts LLM-generated rationales about topic category, audience, and timing into enriched features for a lightweight regressor, bridging contextual insight with numeric precision. Evaluation on the HashView Weibo benchmark demonstrates that reasoning-augmented inputs improve accuracy (RMSE, SRC) across multiple regressors, while also yielding human-readable explanations. The approach offers a scalable, interpretable solution for social media trend forecasting and highlights the value of using LLMs as reasoning engines rather than direct numeric predictors.

Abstract

Hashtag trends ignite campaigns, shift public opinion, and steer millions of dollars in advertising spend, yet forecasting which tag goes viral is elusive. Classical regressors digest surface features but ignore context, while large language models (LLMs) excel at contextual reasoning but misestimate numbers. We present BuzzProphet, a reasoning-augmented hashtag popularity prediction framework that (1) instructs an LLM to articulate a hashtag's topical virality, audience reach, and timing advantage; (2) utilizes these popularity-oriented rationales to enrich the input features; and (3) regresses on these inputs. To facilitate evaluation, we release HashView, a 7,532-hashtag benchmark curated from social media. Across diverse regressor-LLM combinations, BuzzProphet reduces RMSE by up to 2.8% and boosts correlation by 30% over baselines, while producing human-readable rationales. Results demonstrate that using LLMs as context reasoners rather than numeric predictors injects domain insight into tabular models, yielding an interpretable and deployable solution for social media trend forecasting.

Paper Structure

This paper contains 8 sections, 2 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Formulation
  4. Our Approach: BuzzProphet
  5. Experiments
  6. Experimental Setup
  7. Results
  8. Conclusion

Figures (4)

  • Figure 1: Comparison of BuzzProphet with prior work.
  • Figure 2: Domain distribution of our HashView benchmark.
  • Figure 3: Temporal distribution of hashtag postings in HashView, bucketized by hour of day.
  • Figure 4: Illustration of how BuzzProphet generates more accurate predictions through interpretable reasoning. (Orange: LLM predictions for the three dimensions; blue: explanations about their potential influence on hashtag popularity.)