How do you determine the sentiment?

Our sentiment is based on keyword analysis. We currently tag individual news as either Positive, Negative or Neutral. When determining the sentiment we take into account both the title and description. In addition, we look at the importance and frequency of specific keywords.

Note that our system also considers Wall Street Slang keywords such as Bullish, Bearish, Dovish, Hawkish Dead Cat Bounce, Short Squeeze and many more... 

Naturally, it can happen that a specific news item has an incorrect sentiment assigned. However, over time we have enhanced our algorithms and have drastically reduced false positives and negatives. We keep working hard to improve our systems and provide the best data possible :)