extractors.SDF¶

Module: extractors.SDF¶

This program implements the symbolic dynamic filtering (SDF) algorithm as a time series feature extraction method, presented in the following paper:

Bahrampour, Soheil, Asok Ray, Soumalya Sarkar, Thyagaraju Damarla, and Nasser M. Nasrabadi. “Performance comparison of feature extraction algorithms for target detection and classification.” Pattern Recognition Letters 34, no. 16 (2013): 2126-2134.

Written by Soheil Bahrampour, August 2012 Rewritten in Python by Arno Klein, October 2015

Functions¶

mhealthx.extractors.SDF.AnalyzeSymbolSeq(symbolSeq, numStates, morphMatrixFlag)¶

Estimate the Morph matrix and its eigenvector corresponding to eigenvalue 1 by counting.

symbolSeq : numpy array numStates : integer morphMatrixFlag : Boolean

morphMatrix : numpy array PVec : numpy array

>>> import numpy as np
>>> from mhealthx.SDF import AnalyzeSymbolSeq
>>> symbolSeq = np.array([[1,2,3],[4,5,6]])
>>> numStates = 10
>>> morphMatrixFlag = False
>>> morphMatrix, PVec = AnalyzeSymbolSeq(symbolSeq, numStates, morphMatrixFlag)

mhealthx.extractors.SDF.FeatureExtractionSDF(data, partition, numSymbol, PiMatrixFlag)¶

data : numpy array partition : numpy array numSymbol : integer PiMatrixFlag : Boolean

feature : numpy array

>>> import numpy as np
>>> from mhealthx.SDF import FeatureExtractionSDF
>>> data = np.random.random((3,4))
>>> data = data.flatten()
>>> numSymbol = 10
>>> partition = maxEntropyPartition(data, numSymbol)
>>> PiMatrixFlag = False
>>> feature = FeatureExtractionSDF(data, partition, numSymbol, PiMatrixFlag)

mhealthx.extractors.SDF.SDF(trainData, testData)¶

Main function for SDF Feature Extraction.

Each column of trainData and testData contains a time series.

trainData : numpy array testData : numpy array

featuresTrain : numpy array featuresTest : numpy array

>>> import numpy as np
>>> from mhealthx.SDF import SDF
>>> data = np.random.random((3,4))
>>> data = data.flatten()
>>> numSymbol = 10
>>> partition = maxEntropyPartition(data, numSymbol)
>>> PiMatrixFlag = False
>>> featuresTrain, featuresTest = SDF(trainData, testData)

mhealthx.extractors.SDF.SymbolGen(rawData, partition)¶

Generate symbol sequence (symData) of a given time series (rawData) using given partition:

rawData : numpy array partition : numpy array

symData : numpy array

>>> import numpy as np
>>> from mhealthx.SDF import SymbolGen, maxEntropyPartition
>>> rawData = np.random.random((3,4))
>>> rawData = rawData.flatten()
>>> numSymbol = 10
>>> partition = maxEntropyPartition(rawData, numSymbol)
>>> symData = SymbolGen(rawData, partition)

mhealthx.extractors.SDF.maxEntropyPartition(partitionData, numSymbol)¶

Perform maximum entropy partitioning on given data.

partitionData : numpy array numSymbol : integer

partition : numpy array

>>> import numpy as np
>>> from mhealthx.SDF import maxEntropyPartition
>>> partitionData = np.random.random((3,4))
>>> numSymbol = 10
>>> partition = maxEntropyPartition(partitionData, numSymbol)