The Kalman Filter is a time series estimation algorithm based on bayesian statistics. It is a powerful tool for combining information and dealing with uncertainty, and it is widely used to estimate unobservable or unreliable data.
The Kalman Filter is a time series estimation algorithm based on bayesian statistics. It is a powerful tool for combining information and dealing with uncertainty, and it is widely used to estimate unobservable or unreliable data in dinamyc systems. The Kalman Filter is recursive, which allows to process data as it arrives. Despite all its advantages it has small computational requirements.
The Kalman Filter was developed in the 1960´s and played a key role in the development of the Apolo Program that made the Moon Landing possible. Since then it has grown in importance and it is now widely used in robotics, guiding systems, computer design and more recently in finance.
The Kalman Filter processes information from two different sources, measurement and state equations. It weights the quality and quantity of the information provided by these two sources, and then it combines both optimally to create a brand new set of information that does not depend on missing or faulty data.
In finance it is very common to deal with data problems, very often the data is not fully available or it is unreliable. The Kalman filter allows us to recreate data sets of better quality than the originals.
The filter will “run over” any gap in the data series filling it, it will also identify and correct erroneous data. Potentially it could be used to recreate data series that are otherwise hard to find in the markets.
If you want to know more about Kalman Filter, you could view the complete article here: