Discrete wavelet transform
The Discrete wavelet transform indicator is an enhanced implementation of the numerical and functional analysis for MetaTrader 4, the discrete wavelet transform method (DWT). The specific feature of DWT is the way the wavelet is represented in the form of a discrete signal (sample). DWT is widely used for filtering and preliminary processing of data, for analyzing the state and making forecasts for the stock and currency markets, for recognizing patterns, processing and synthesizing various signals, for example, speech and medical signals, for solving the problems of image compression and processing, in learning neural networks and in many other cases.
The Discrete wavelet transform indicator is a filter, which is used for finding optimal market entry and exit points in trending conditions. It allows for more accurate forecasting and analysis of the current and emerging trend.Discrete wavelet transform includes several methods for filtering the analyzed signal.
- 1. Haar (haar)
- 2. Daubechies (db1, db2, db3, db4, db5, db6, db7, db8, db9, db10)
- 3. Coiflet (coif1, coif2, coif3, coif4, coif5)
- 4. Biorthogonal (bior11, bior13, bior15, bior22, bior24 и др)
- 5. Reverse Biorthogonal (rbior11, rbior13, rbior15, rbior22, bior24 и др)
as well as the calculation options and final filter presentation variants
- 1. DWT is discrete wavelet transform
- 2. MODWT is discrete wavelet transform with overlap
- 3. SWT is stationary wavelet transform
Filters differ in basic functions, wavelet transform coefficients, as well as calculation methods (mainly iterative). Each filter has its own characteristics in the time and frequency space. The use of different filters allows identifying certain properties of the analyzed signal.
The number of wavelets used in signal analysis determines the level of decomposition. The accuracy of the signal presentation decreases on higher levels of decomposition, but the possibility of filtering the signal at a higher quality with better noise removal and efficient compression increases.
- type - calculation type (0-DWT, 1-MODWT, 2-SWT);
- filter - filter type;
- indicator - indicator used for transform (0-1);
- length - indicator length;
- J - the level of decomposition.
2. Added new filter type - Reverse Biorthogonal (rbior)
3. Added two new indicators for transform
4. Extended filter families (filter order)
- Daubechies - db1, db2, db3, db4, db5, db6, db7, db8, db9, db10;
- Coiflet - coif1, coif2, coif3, coif4, coif5;
- Biorthogonal - bior11, bior13, bior15, bior22, bior24, bior31, bior33, bior35, bior37, bior39, bior44, bior55, bior68;
- Reverse Biorthogonal - rbior11, rbior13, rbior15, rbior22, rbior24, rbior31, rbior33, rbior35, rbior37, rbior39, rbior44, rbior55, rbior68.
5. Extended the range of possible values for the time series length parameter