Discussion of article "Using the Kalman Filter for price direction prediction"

New comment
 

New article Using the Kalman Filter for price direction prediction has been published:

For successful trading, we almost always need indicators that can separate the main price movement from noise fluctuations. In this article, we consider one of the most promising digital filters, the Kalman filter. The article provides the description of how to draw and use the filter.

Currency and stock charts feature price fluctuations having different frequency and amplitude. Our task is to determine the main trends based on these short and long movements. Some traders draw trendlines on the chart, others use indicators. In both cases, our purpose is to separate the true price movement from noise caused by the influence of minor factors that have a short-term effect on the price. In this article I propose using the Kalman filter to separate the major movement from the market noise.

Author: Dmitriy Gizlyk

 
Hi, I looked at your Kalman filter indicator and I think it is very good. However, it does not plot on cash indices-dow jones, DAX, S&P500. I think the problem is that with the indices tick != point. Also, When I try to compile the Kalman indy, I get the following error:
cannot cast 'L1' to 'D1' Math.mqh 20 30
Could you please fix the indicator? I think it has a lot of value!
Kind regards,
Stanisav
 


this is the test result with conditions of article but the date is only from 01-8-2017 to 30-10-2017 (instead of 30-8-2017 in article)

 
Hello, your idea is to fit the data and then extrapolate.
But I want to get the fitted curve data, what should I do?  How is the code written?
Just like regression analysis, the fitted data can be extrapolated. Also we can get the fitted curve data.
New comment