Can investors’ emotions on the stock exchange be measured? Absolutely! Various methods exist, from surveys like AAII Bull and Bear, to indices such as Fear and Greed, and analysis of buying trends. But can emotions themselves be directly measured? Let’s explore this in the article below.
Sentiment Analysis in Stock Market Forecasting
market indices
In December 2018, the National Bureau of Economic Research published “Media Sentiment and International Asset Prices,” analyzing Reuters and Bloomberg articles’ impact on stock indices. Key findings include:
- News sentiment correlates more strongly with bear markets.
- Sentiment is a better predictor of global indices than VIX (CBOE Volatility Index).
- Positive media sentiment influences developed markets positively but affects emerging markets negatively.
Unfortunately, these methods only achieve 55-60% accuracy, likely because they rely on emotional tags from social media.
Emotional monitoring for stock price prediction
What if we combined media and emotion monitoring using deep neural networks? Sentimenti’s tools allow for the analysis of 11 emotional variables in texts (8 basic emotions). We track sentiment and specific emotions, measuring their intensity and changes over time. We then compare these findings with hard indicators, like the WIG-Banks index and company share prices.
Emotions on the Warsaw Stock Exchange
In late 2018, we tested our models with data from 50 Polish companies. Remarkably, in 87.1% of cases, changes in emotional intensity predicted stock price changes! Our report, featuring companies like CD Projekt and KGHM, details these findings.
Conclusion
Monitoring emotions can be a valuable tool for investors, translating market sentiment into actionable data without the need for complex surveys. We’ve also applied this to the cryptocurrency market with similar success.