A step-by-step look at how the Sigmita Sentiment Engine reads headlines, shows their age and scores and turns them into one bearish-to-bullish reading per market.
A Sigmita sentiment reading condenses a day of market news into one number you can check. This guide explains how the Sigmita Sentiment Engine works, step by step, and how to read what it shows.
What is a market sentiment score?
A sentiment score summarises whether recent news about a market leans bullish (supportive of higher prices), bearish (supportive of lower prices) or neutral. Sigmita scores every headline on a scale from bearish to bullish, then combines the scores into one reading per market.
How does Sigmita build a reading?
- Collect: headlines for each market are gathered continuously.
- Score: the Sigmita Sentiment Engine, our own finance-tuned language model, scores each headline from bearish to bullish.
- Weigh: each headline carries a reliability weight, so one noisy feed cannot dominate the average.
- Combine: scores roll up into a single sentiment value for the market, shown beside macro themes such as central-bank policy and the US dollar.
- Explain: the headlines behind the score are listed with their age and individual score.
What do confidence and top catalyst mean?
Confidence shows how clearly the headlines point in one direction. When confidence is low, the reading says so and the platform flags caution instead of implying certainty. The top catalyst is the headline that moved the reading most, so you can open it and judge it yourself.
What are the limits of a sentiment score?
- News sentiment describes what is being reported, not what prices will do.
- A reading can lag fast-moving events and can be thin when few articles are available.
- It is research context for your own judgement, never a recommendation to buy or sell.