Scholarlink

This study introduces a novel approach to stock market prediction by integrating LinkedIn job postings with traditional financial and macroeconomic indicators. While existing models rely on historical price data and lagging metrics, our framework leverages real-time hiring trends as forward-looking signals of corporate strategy and sector shifts. We apply natural language processing to extract skill demand patterns from job descriptions and combine them with structured financial data to uncover predictive correlations between labor market activity and stock returns in the technology sector. Our Random Forest model, trained on a hybrid dataset, achieves 37% explanatory power ( \(R^2\) ) for 30-day forward returns. Job posting views (44.9%) and application rates (41.8%) emerge as key predictors. Despite limitations in ticker matching and sector scope, our results show that labor market activity provides an early signal of stock performance. This work bridges labor economics and financial forecasting, offering a replicable, scalable framework for anticipating market trends using alternative data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data-Driven Social Signal Mining for Stock Return Modeling via LinkedIn Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Nick Yennerell,
  • Benjamin Scott,
  • Luke Rogers,
  • Aaron March,
  • Behnaz Moradi-Jamei

摘要

This study introduces a novel approach to stock market prediction by integrating LinkedIn job postings with traditional financial and macroeconomic indicators. While existing models rely on historical price data and lagging metrics, our framework leverages real-time hiring trends as forward-looking signals of corporate strategy and sector shifts. We apply natural language processing to extract skill demand patterns from job descriptions and combine them with structured financial data to uncover predictive correlations between labor market activity and stock returns in the technology sector. Our Random Forest model, trained on a hybrid dataset, achieves 37% explanatory power ( \(R^2\) ) for 30-day forward returns. Job posting views (44.9%) and application rates (41.8%) emerge as key predictors. Despite limitations in ticker matching and sector scope, our results show that labor market activity provides an early signal of stock performance. This work bridges labor economics and financial forecasting, offering a replicable, scalable framework for anticipating market trends using alternative data.