LSEG Labs Project: Mosaic
Explaining extreme price moves with machine learning
Extreme price movements occur suddenly and without warning, often affecting pricing and volatility.
Traders typically go through a slow and painstaking process of sourcing and analysing disparate information to identify the cause of a price anomaly.
Financial institutions have a significant amount and variety of datasets at their disposal, but these are rarely combined to uncover new insights or explain events.
Mosaic combines real-time analytics, machine learning, and intelligent search processes across different datasets to explain sudden price moves.
Real-time analytics are used to detect price movement, while an anomaly detection machine learning model, trained using historical data, confirms that a price move is anomalous.
Mosaic connects relevant pricing, news, social media and events data, and shows individual news stories, social media comments, publication times and events as nodes on the price graph. This provides direct correlation between the stock price and factors that have affected it.
It also tracks price movements related to other companies in the sector to identify relevant industry trends.
Mosaic in action
The Mosaic prototype includes:
- Real-time identification of anomalous price moves in the market.
- An alert summary panel to help provide context to the price move by listing historical price summary information and highlighting the most important information associated with the move across events, news and sectors.
- Provision of analysis for both real-time and historical price moves
What we’re thinking next
LSEG Labs are taking the experience of creating advanced analytics for anomaly detection and the explanation of price moves forward and developing prototypes that arm traders with new insight to quickly assess a price move or event, and make decisions.