(SMA) S-Factor Sentiment

Leverage Social Market Analytics’ patented technology to identify and aggregate market-moving tweets and other social sentiment data. Independent research has shown that SMA’s social media-based factors are predictive at statistically significant levels and can provide alpha enhancement and risk reduction to most quantitative models.

Incorporate SMA data into multi factor Quantitative models, Algo Execution, Market Making, Indices, Barra and VarRisk models and Research. Predictive signals range in frequency from real-time to quarterly. Data for sectors and industries is aggregated from the security level. SMA has point in time out of sample sentiment data dating back to 2011 for accurate backtests.

Identify professional investors with SMA proprietary algorithms. SMA’s process is driven by two U.S. patents that extract, evaluate, and calculate nearly 1B textual sentiment based data points per day to provide real time APIs. SMA’s financial machine learning NLP has been built over the past seven years using supervised and unsupervised training. All SMA data is out-of-sample and all changes are applied on a go forward basis.

*Powered by FactSet

搜集汇总

背景与挑战

背景概述

该数据集采用专利算法分析每日近10亿条社交媒体文本，生成可回溯至2011年的样本外情感因子数据，适用于量化模型构建与回测。其独特之处在于通过金融机器学习NLP技术，将实时至季度频率的预测信号聚合到行业层面，显著提升模型表现。