Bloomberg whitepapers - Sentiment
Twitter Finance Analytics
Bloomberg seminars and extra
Bloomberg technical papers
We have found that sentiment as reflected in news stories in the Bloomberg End-of-Day News and Analytics File has an impact on the prices of large-cap U.S. stocks, with the impact differing across stories depending on the set of topic codes with which they are tagged
When rational arbitrageurs have limited risk-bearing capacity and time horizons, the actions of irrational noise traders can affect asset prices (De Long, Shleifer, Summers, & Waldmann, 1990a). Such actions can be interpreted as being driven by fluctuating investor sentiment. This creates the possibility of trading profitably on the basis of investor sentiment, most obviously by being a contrarian, but, under some circumstances, it may be rational to “jump on the bandwagon” and bet with, rather than against, noise traders (De Long, Shleifer, Summers, & Waldmann, 1990b). Various proxies for investor sentiment have been proposed (Baker & Wurgler, 2006), but perhaps the most direct way to measure sentiment in the stock market is to analyze the words of those who are commenting on stocks. One traditional source of such comments is stories in the news media (Tetlock, 2007).
More recently, Google searches and Twitter feeds have been used (Mao, Counts, & Bollen, 2015).
Sentiment impact on stock prices of news with selected topic codes: Part Two Ivailo Dimov - 2018
We use Latent Semantic Analysis (LSA) with a suitable Independent Component Analysis (ICA)
regularization to retrieve latent, interpretable topic code factors in Bloomberg’s machine-readable
three types of equity trading strategies based on sentiment data
Lei Huang Newsfeed dymistified
... paint a coherent picture: the divergence is striking between what people read and what moves the markets
IDEA - ... CONSTRUCT NEWS-DRIVEN fundamental context critical in the determination of longer-term price movements