Filtering (FSharp.Stats) | FSharp.Stats

FSharp.Stats

Filtering Module

Namespace: FSharp.Stats.Signal
Assembly: FSharp.Stats.dll

Functions and values

Function or value	Description
`correlateValid kernel signal` Full Usage: `correlateValid kernel signal` Parameters: kernel : `Vector<float>` - Vector of length K (FIR filter, etc.). signal : `Vector<float>` - Vector of length N. Returns: `float[]` An array of length max(1, N-K+1) with the "valid" correlation result. Typically used for FIR filtering.	"Valid" cross-correlation of a kernel kernel with a signal signal. Returns an array of length (signal.Length - kernel.Length + 1) if signal.Length >= kernel.Length, otherwise a single dot product if kernel.Length > signal.Length. kernel : `Vector<float>` Vector of length K (FIR filter, etc.). signal : `Vector<float>` Vector of length N. Returns: `float[]` An array of length max(1, N-K+1) with the "valid" correlation result. Typically used for FIR filtering.
`optimizeWindowWidth polOrder windowWidthToTest blankSignal signalOfInterest` Full Usage: `optimizeWindowWidth polOrder windowWidthToTest blankSignal signalOfInterest` Parameters: polOrder : `int` - windowWidthToTest : `int[]` - blankSignal : `float[]` - signalOfInterest : `float[]` - Returns: `int`	Estimates the autocorrelation at lag 1 of a blank signal (containing only noise). Subsequently, the signal of interest is smoothed several times by a savitzky golay filter using constant polynomial order and variing windowWidth. For each iteration, the deviation of the smoothed to the original signal is computed and the autocorrelation at lag 1 of this residual noise is computed. The function returns the optimized window width yielding a autocorrelation at lag 1 closest to the value computed for the blank signal. Method is based on: https://doi.org/10.1021/ac0600196 polOrder : `int` windowWidthToTest : `int[]` blankSignal : `float[]` signalOfInterest : `float[]` Returns: `int` Example
`savitzkyGolay windowSize order deriv rate data` Full Usage: `savitzkyGolay windowSize order deriv rate data` Parameters: windowSize : `int` - Must be odd and at least (order+2) if deriving. order : `int` - Polynomial order (must be >= deriv). A higher order can better fit curvature. deriv : `int` - Order of derivative to compute (0 => smoothing). rate : `float` - Scaling factor for derivative, typically sampling rate = 1.0 if no special scale needed. data : `Vector<float>` - The data vector to filter (length N). Returns: `float[]` A float[] array with the filtered (or derived) signal, of length = data.Length.	Smooth (or differentiate) data with a Savitzky–Golay filter of given windowSize and polynomial order. The deriv parameter specifies the derivative order to compute (0 = smoothing only). The rate helps scale the derivative if deriv > 0. windowSize : `int` Must be odd and at least (order+2) if deriving. order : `int` Polynomial order (must be >= deriv). A higher order can better fit curvature. deriv : `int` Order of derivative to compute (0 => smoothing). rate : `float` Scaling factor for derivative, typically sampling rate = 1.0 if no special scale needed. data : `Vector<float>` The data vector to filter (length N). Returns: `float[]` A float[] array with the filtered (or derived) signal, of length = data.Length.