Scholarlink

<p>This study examines the influence of emojis, recognized as a distinct form of communication, on the efficient market hypothesis (EMH) within the Bitcoin market over a 12-year period. Using a methodology that integrates a comprehensive analysis with 6 decision tree models, SHapley Additive exPlanation feature selection, and ridge regression, we examine an extensive dataset of 3,733 emojis from 136,546,837 tweets. Our best performance model, featuring 195(219), 192(210), 93(51), and 93(54) emojis identified in original tweets as indicators of positive(negative) influence, predicts probabilities of 62.80% (61.17%), 77.87% (75.71%), 84.39% (82.26%), and 85.31% (83.36%), respectively, for succeeding market ascents (descents) across 1-, 6-, 12-, and 24-hour periods. These outcomes substantiate the semistrong version of the EMH in the Bitcoin market context.</p>

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

Smiles or tears? How emojis power a semistrong efficient market of bitcoin: A comprehensive analysis of decision tree models

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月14日

查看全文
  • Fang Wang,
  • Adriana F. Gabor,
  • Marko Gacesa,
  • Ernesto Damiani

摘要

This study examines the influence of emojis, recognized as a distinct form of communication, on the efficient market hypothesis (EMH) within the Bitcoin market over a 12-year period. Using a methodology that integrates a comprehensive analysis with 6 decision tree models, SHapley Additive exPlanation feature selection, and ridge regression, we examine an extensive dataset of 3,733 emojis from 136,546,837 tweets. Our best performance model, featuring 195(219), 192(210), 93(51), and 93(54) emojis identified in original tweets as indicators of positive(negative) influence, predicts probabilities of 62.80% (61.17%), 77.87% (75.71%), 84.39% (82.26%), and 85.31% (83.36%), respectively, for succeeding market ascents (descents) across 1-, 6-, 12-, and 24-hour periods. These outcomes substantiate the semistrong version of the EMH in the Bitcoin market context.