StockEmotions

该数据集是一个专门用于检测股市情绪的集合，包含了从金融社交媒体平台StockTwits收集的1万个英文评论。这些评论被细分为12种细致的情绪类别，旨在分析新冠疫情期间投资者的情绪。此外，该数据集还包括了两类金融情绪（看涨和看跌）以及十二种情绪类别。为了确保公平代表并减少用户偏见，数据集经过了精心筛选，并包含了对于情绪检测至关重要的带表情符号的评论。规模上，该数据集涵盖了1万个评论，覆盖了标普500指数超过80%的市场资本化。其任务是对金融情绪/情绪进行分类以及进行多变量时间序列预测。

This dataset is a collection specifically designed for stock market sentiment detection. It contains 10,000 English comments collected from the financial social media platform StockTwits. These comments are categorized into 12 fine-grained sentiment categories, aiming to analyze investors' sentiment during the COVID-19 pandemic. Additionally, the dataset covers two broad financial sentiment types (bullish and bearish) alongside the 12 fine-grained sentiment categories mentioned above. To ensure fair representation and reduce user bias, the dataset has been carefully curated, and includes comments with emojis which are critical for sentiment detection. In terms of scale, the dataset encompasses 10,000 comments, representing over 80% of the total market capitalization of the S&P 500 index. The supported tasks include financial sentiment classification and multivariate time series forecasting.

搜集汇总

背景与挑战

背景概述

StockEmotions是一个金融领域数据集，专注于情感分析和多变量时间序列预测。它包含2020年收集的10,000条样本，标注了2种情感（看涨和看跌）和12种情绪类别，并附带38家公司的历史价格数据，适用于金融情感分类和股票市场预测任务。