Get started

The purpose of this package is to make tabular data from ECG-recordings by calculating features. The package is built on WFDB 1 and NeuroKit2 2. The package can be convenient when doing machine learning on ECG-data.

Usage:

Featurize .dat-files:

from ECGfeaturizer import featurize as ef

# Make ECG-featurizer object
Feature_object =ef.get_features()

# Preprocess the data (filter, find peaks, etc.)
My_features=Feature_object.featurizer_dat(features=ecg_filenames,labels=labels,directory="./data/",demographical_data=demo_data)

Featurize .mat-files:

from ECGfeaturizer import featurize as ef

number_of_ECGs = <the amount of ECGs>
directory = "<your dir>"

# Make ECG-featurizer object
Feature_object =ef.get_features()

# Preprocess the data (filter, find peaks, etc.)
My_features=Feature_object.featurizer_mat(num_features=number_of_ECGs, mat_dir = directory)

Input data:

features:

A numpy array of ECG-recordings in directory. Each recording should have a file with the recording as a time series and one file with meta data containing information about the patient and measurement information. This is standard format for WFDB and PhysioNet-files 1 3

Supported input files:

Input data

Supported file format

ECG-recordings

.dat files

Patient meta data

.hea files
labels:

A numpy array of labels / diagnoses for each ECG-recording. The length of the labels-array should have the same length as the features-array

len(labels) == len(features)
directory:

A string with the path to the features. If the folder structure looks like this:

mypath
├── ECG-recordings
│ ├── A0001.hea
│ ├── A0001.dat
│ ├── A0002.hea
│ ├── A0002.dat
│ └── Axxxx.dat

then the feature and directory varaible could be:

demographical_data:

The demographical data that is used in this function is age and gender. A Dataframe with the following 3 columns should be passed to the featurizer() function.

age

gender

filename_hr

0

11.0

1

“A0001”

1

57.0

0

“A0002”

2

94.0

0

“A0003”

3

34.0

1

“A0004”

The strings in the filename_hr -column should be the same as the strings in the feature array. In this example gender is OneHot encoded such that .. math:

1 = Female
0 = Male

Output Features:

The features that are calculated in present version is:

Index number

Feature Name

Description

0

gender

Patients Gender

1

age

Patients Age

2

R HR STD

Heart rate standard deviation derived from R-peaks

3

R HR median

Heart rate median derived from R-peaks

4

R HR min

Heart rate minimum derived from R-peaks

5

R HR max

Heart rate maximum derived from R-peaks

6

R HR mean

Heart rate mean derived from R-peaks

7

RMSSD

Heart rate variability calculated using RMSSD

8

R amp II std

Standard deviation of R-peak amplitude in lead II

9

R amp II min

Minimum R-peak amplitude in lead II

10

R amp II min_2

this was supposed to be R amp II max

11

R amp leads I

Voltage amplitude of R-peak in lead I

12

R amp leads II

Voltage amplitude of R-peak in lead II

13

R amp lead III

Voltage amplitude of R-peak in lead III

14

R amp lead aVR

Voltage amplitude of R-peak in lead aVR

15

R amp lead aVL

Voltage amplitude of R-peak in lead aVL

16

R amp lead aVF

Voltage amplitude of R-peak in lead aVF

17

R amp V1

Voltage amplitude of R-peak in lead V1

18

R amp V2

Voltage amplitude of R-peak in lead V2

19

R amp V3

Voltage amplitude of R-peak in lead V3

20

R amp V4

Voltage amplitude of R-peak in lead V4

21

R amp V5

Voltage amplitude of R-peak in lead V5

22

R amp V6

Voltage amplitude of R-peak in lead V6

23

p_offset_std

Standard deviation of heart rate calculated from P-offset

24

p_offset_median

Median heart rate calculated from P-offset

25

p_offset_min

Minimum heart rate calculated from P-offset

26

p_offset_max

Maximum heart rate calculated from P-offset

27

mean_p_offset

Mean heart rate calculated from P-offset

28

p_onsets_std

Standard deviation of heart rate calculated from P-onset

29

p_onsets_median

Median heart rate calculated from P-onset

30

p_onsets_min

Minimum heart rate calculated from P-onset

31

p_onsets_max

Maximum heart rate calculated from P-onset

32

mean_p_onsets

Mean heart rate calculated from P-onset

33

ECG_baseline

ECG baseline calculated taking the mean of all P-onset voltages

34

p_rate_std

Standard deviation of heart rate calculated from P-peak

35

p_rate_median

Median heart rate calculated from P-peak

36

p_rate_min

Minimum heart rate calculated from P-peak

37

p_rate_max

Maximum heart rate calculated from P-peak

38

mean_p_rate

Mean heart rate calculated from P-peak

39

P amp leads I

Voltage amplitude of P-peak in lead I

40

P amp leads II

Voltage amplitude of P-peak in lead II

41

P amp lead III

Voltage amplitude of P-peak in lead III

42

P amp lead aVR

Voltage amplitude of P-peak in lead aVR

43

P amp lead aVL

Voltage amplitude of P-peak in lead aVL

44

P amp lead aVF

Voltage amplitude of P-peak in lead aVF

45

P amp V1

Voltage amplitude of P-peak in lead V1

46

P amp V2

Voltage amplitude of P-peak in lead V2

47

P amp V3

Voltage amplitude of P-peak in lead V3

48

P amp V4

Voltage amplitude of P-peak in lead V4

49

P amp V5

Voltage amplitude of P-peak in lead V5

50

P amp V6

Voltage amplitude of P-peak in lead V6

51

q_rate_std

Standard deviation of heart rate calculated from Q-peak

52

q_rate_median

Median heart rate calculated from Q-peak

53

q_rate_min

Minimum heart rate calculated from Q-peak

54

q_rate_max

Maximum heart rate calculated from Q-peak

55

mean_q_rate

Mean heart rate calculated from Q-peak

56

Q amp leads I

Voltage amplitude of Q-peak in lead I

57

Q amp leads II

Voltage amplitude of Q-peak in lead II

58

Q amp lead III

Voltage amplitude of Q-peak in lead III

59

Q amp lead aVR

Voltage amplitude of Q-peak in lead aVR

60

Q amp lead aVL

Voltage amplitude of Q-peak in lead aVL

61

Q amp lead aVF

Voltage amplitude of Q-peak in lead aVF

62

Q amp V1

Voltage amplitude of Q-peak in lead V1

63

Q amp V2

Voltage amplitude of Q-peak in lead V2

64

Q amp V3

Voltage amplitude of Q-peak in lead V3

65

Q amp V4

Voltage amplitude of Q-peak in lead V4

66

Q amp V5

Voltage amplitude of Q-peak in lead V5

67

Q amp V6

Voltage amplitude of Q-peak in lead V6

68

s_rate_std

Standard deviation of heart rate calculated from S-peak

69

s_rate_median

Median heart rate calculated from S-peak

70

s_rate_min

Minimum heart rate calculated from S-peak

71

s_rate_max

Maximum heart rate calculated from S-peak

72

mean_s_rate

Mean heart rate calculated from S-peak

73

S amp leads I

Voltage amplitude of S-peak in lead I

74

S amp leads II

Voltage amplitude of S-peak in lead II

75

S amp lead III

Voltage amplitude of S-peak in lead III

76

S amp lead aVR

Voltage amplitude of S-peak in lead aVR

77

S amp lead aVL

Voltage amplitude of S-peak in lead aVL

78

S amp lead aVF

Voltage amplitude of S-peak in lead aVF

79

S amp V1

Voltage amplitude of S-peak in lead V1

80

S amp V2

Voltage amplitude of S-peak in lead V2

81

S amp V3

Voltage amplitude of S-peak in lead V3

82

S amp V4

Voltage amplitude of S-peak in lead V4

83

S amp V5

Voltage amplitude of S-peak in lead V5

84

S amp V6

Voltage amplitude of S-peak in lead V6

85

t_rate_std

Standard deviation of heart rate calculated from T-peak

86

t_rate_median

Median heart rate calculated from T-peak

87

t_rate_min

Minimum heart rate calculated from T-peak

88

t_rate_max

Maximum heart rate calculated from T-peak

89

mean_t_rate

Mean heart rate calculated from T-peak

90

T amp leads I

Voltage amplitude of T-peak in lead I

91

T amp leads II

Voltage amplitude of T-peak in lead II

92

T amp lead III

Voltage amplitude of T-peak in lead III

93

T amp lead aVR

Voltage amplitude of T-peak in lead aVR

94

T amp lead aVL

Voltage amplitude of T-peak in lead aVL

95

T amp lead aVF

Voltage amplitude of T-peak in lead aVF

96

T amp V1

Voltage amplitude of T-peak in lead V1

97

T amp V2

Voltage amplitude of T-peak in lead V2

98

T amp V3

Voltage amplitude of T-peak in lead V3

99

T amp V4

Voltage amplitude of T-peak in lead V4

100

T amp V5

Voltage amplitude of T-peak in lead V5

101

T amp V6

Voltage amplitude of T-peak in lead V6

102

t_offset_std

Standard deviation of heart rate calculated from T-offset

103

t_offset_median

Median heart rate calculated from T-offset

104

t_offset_min

Minimum heart rate calculated from T-offset

105

t_offset_max

Maximum heart rate calculated from T-offset

106

mean_t_offset

Mean heart rate calculated from T-offset

107

t_onsets_std

Standard deviation of heart rate calculated from T-onset

108

t_onsets_median

Median heart rate calculated from T-onset

109

t_onsets_min

Minimum heart rate calculated from T-onset

110

t_onsets_max

Maximum heart rate calculated from T-onset

111

mean_t_onsets

Mean heart rate calculated from T-onset

References:

1

WFDB: https://github.com/MIT-LCP/wfdb-python

2

Makowski, D., Pham, T., Lau, Z. J., Brammer, J. C., Lesspinasse, F., Pham, H., Schölzel, C., & S H Chen, A. (2020). NeuroKit2: A Python Toolbox for Neurophysiological Signal Processing. Retrieved March 28, 2020, from https://github.com/neuropsychology/NeuroKit

3

Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, Mietus JE, Moody GB, Peng CK, Stanley HE. PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals. Circulation 101(23):e215-e220 [Circulation Electronic Pages; http://circ.ahajournals.org/content/101/23/e215.full]; 2000 (June 13). PMID: 10851218; doi: 10.1161/01.CIR.101.23.e215