feature: add G-H filter and Kalman filter #915
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov Report
@@ Coverage Diff @@
## main #915 +/- ##
==========================================
+ Coverage 19.71% 19.84% +0.12%
==========================================
Files 398 402 +4
Lines 29950 30049 +99
==========================================
+ Hits 5904 5962 +58
- Misses 23473 23510 +37
- Partials 573 577 +4
Continue to review full report at Codecov.
|
- implement G-H (alpha beta) filter and Kalman filter - compare the predict accurateness with other indicator
COLDTURNIP
force-pushed
the
feature/observer_filters
branch
from
0a0dd89 to
9c684c1
Compare
COLDTURNIP
commented
Sep 4, 2022
| for _, tt := range tests { | ||
| t.Run(tt.name, func(t *testing.T) { | ||
| filter := &GHFilter{IntervalWindow: types.IntervalWindow{Window: tt.args.window}} | ||
| ewma := &EWMA{IntervalWindow: types.IntervalWindow{Window: tt.args.window}} |
There was a problem hiding this comment.
EWMA is widely used to calculate average price with less latency compare to SMA. Do we expect the comparison of ZLEMA?
| for _, tt := range tests { | ||
| t.Run(tt.name, func(t *testing.T) { | ||
| filter := &KalmanFilter{IntervalWindow: types.IntervalWindow{Window: tt.args.window}} | ||
| ewma := &EWMA{IntervalWindow: types.IntervalWindow{Window: tt.args.window}} |
There was a problem hiding this comment.
Same here, change to ZLEMA?
austin362667
approved these changes
Sep 4, 2022
c9s
reviewed
Sep 5, 2022
c9s
reviewed
Sep 5, 2022
c9s
reviewed
Sep 5, 2022
c9s
reviewed
Sep 5, 2022
|
You can add these two indicators to the standardIndicator struct |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Observer filters are realtime estimation algorithms to measure the status changes in a noisy environment. These recursive filtering algorithms estimate the true target states by minimize the uncertainties from system noise and observation noise. Such technologies are useful in financial econometrics as estimators without any latency, and more accurate than moving average based indicators.
The G-H filter in this PR implement John Ehlers' work. The system uncertainty is defined as the change of middle prices of K-lines, and the observation uncertainty is the price range of each K-line. Instead of fixed alpha & beta of conventional G-H filter, these two factor are calculated using an exponential recursive method controlled by configured window length.
More generalized than G-H filter, Kalman filter replaces these update factors by an adaptive factor Kalman gain, calculates the gain from uncertainties of estimation and observation, and optimize the estimation base on the gain, the measurement, and the previous prediction. In our implementation, we add an additional factor
AdditionalSmoothWindowin need of smoothness, and set to 0 (classic Kalman filter) by default.Estimation Accurateness
In this PR, tests are created to compare the estimation accurateness between the two filters and EMA by calculating the square errors.
References