def get_global_feature(self):
"""
获取时域全局特征，包含最大值、标准差、平均值
:param hadcropped:
:return:
"""
square_data, square_energy, square_azrate = self.pre_process(method='hanning', ifcrop=True)
func = lambda x: [
# feature_calc.autocorrelation(norm(x), 5),
np.std(x),
feature_calc.approximate_entropy(norm(x), 5, 1),
feature_calc.cid_ce(x, normalize=True),
feature_calc.count_above_mean(x),
feature_calc.first_location_of_minimum(x),
feature_calc.first_location_of_maximum(x),
feature_calc.last_location_of_maximum(x),
feature_calc.last_location_of_minimum(x),
feature_calc.longest_strike_above_mean(x),
feature_calc.number_crossing_m(x, 0.8*np.max(x)),
feature_calc.skewness(x),
feature_calc.time_reversal_asymmetry_statistic(x, 5)
]
# global features I want to get
upper_rate = self.get_upper_rate(square_energy)
feature = np.hstack([
[np.mean(norm(square_energy))],
[upper_rate],
func(square_azrate),
func(square_energy)
])
return feature
def complexity(mag):
"""This function calculator is an estimate for a time series complexity.
A higher value represents more complexity (more peaks,valleys,etc.)
See: Batista, Gustavo EAPA, et al (2014). CID: an efficient complexity-invariant
distance for time series. Data Mining and Knowledge Difscovery 28.3 (2014): 634-669.
rtype: float
"""
c = ts.cid_ce(mag, True)
return c
def get_mfcc_feature(self, hadcropped=False):
'''
calculate Mel-frequency cepstral coefficients in frequency domain and extract features from MFCC
:return: numpy array
'''
assert self.frame_per_second not in [32, 64, 128, 256], \
Exception("Cannot operate butterfly computation ,"
"frame per second should in [32, 64, 128, 256]")
hanning_kernel = self.get_window(method='hanning')
windowed = self._add_window(hanning_kernel, self.meta_audio_data) # [num_frame, kernel_size]
hanning_energy = self.get_energy(self.meta_audio_data, hanning_kernel)
if not hadcropped:
boundary = self.get_boundary(hanning_energy)
cropped = windowed[boundary[0]: boundary[1] + 1, :]
frequency = np.vstack([fft.fft(frame.squeeze()) for frame in np.vsplit(cropped, len(cropped))])
else:
frequency = np.vstack([fft.fft(windowed)])
frequency = np.abs(frequency)
frequency_energy = frequency ** 2
low_freq = self.sr / self.num_per_frame
high_freq = self.sr
H = self._mfcc_filter(self.mfcc_cof, low_freq, high_freq)
S = np.dot(frequency_energy, H.transpose()) # (F, M)
cos_ary = self._discrete_cosine_transform()
mfcc_raw_features = np.sqrt(2 / self.mfcc_cof) * np.dot(S, cos_ary) # （F，N)
upper = [self.get_upper_rate(fea) for fea in mfcc_raw_features.transpose()]
assert len(upper) == mfcc_raw_features.shape[1]
func = lambda x: [
# feature_calc.autocorrelation(norm(x), 5),
np.std(x),
feature_calc.approximate_entropy(norm(x), 5, 1),
feature_calc.cid_ce(x, normalize=True),
feature_calc.count_above_mean(x),
feature_calc.first_location_of_minimum(x),
feature_calc.first_location_of_maximum(x),
feature_calc.last_location_of_maximum(x),
feature_calc.last_location_of_minimum(x),
feature_calc.longest_strike_above_mean(x),
feature_calc.number_crossing_m(x, 0.8*np.max(x)),
feature_calc.skewness(x),
feature_calc.time_reversal_asymmetry_statistic(x, 5)
]
mfcc_features = np.hstack(
[func(col) for col in mfcc_raw_features.transpose()]
)
return mfcc_features
def features(self, x, y, seg_id):
feature_dict = dict()
feature_dict['target'] = y
feature_dict['seg_id'] = seg_id
x = pd.Series(denoise_signal(x, wavelet='db1', level=1))
#x = x - np.mean(x)
zc = np.fft.fft(x)
zc = zc[:37500]
# FFT transform values
realFFT = np.real(zc)
imagFFT = np.imag(zc)
freq_bands = [x for x in range(0, 37500, 7500)]
magFFT = np.sqrt(realFFT**2 + imagFFT**2)
phzFFT = np.arctan(imagFFT / realFFT)
phzFFT[phzFFT == -np.inf] = -np.pi / 2.0
phzFFT[phzFFT == np.inf] = np.pi / 2.0
phzFFT = np.nan_to_num(phzFFT)
for freq in freq_bands:
if freq == 0:
continue
feature_dict['FFT_Mag_01q%d' % freq] = np.quantile(
magFFT[freq:freq + 7500], 0.01)
feature_dict['FFT_Mag_10q%d' % freq] = np.quantile(
magFFT[freq:freq + 7500], 0.1)
feature_dict['FFT_Mag_90q%d' % freq] = np.quantile(
magFFT[freq:freq + 7500], 0.9)
feature_dict['FFT_Mag_99q%d' % freq] = np.quantile(
magFFT[freq:freq + 7500], 0.99)
feature_dict['FFT_Mag_mean%d' % freq] = np.mean(magFFT[freq:freq +
7500])
feature_dict['FFT_Mag_std%d' % freq] = np.std(magFFT[freq:freq +
7500])
feature_dict['FFT_Mag_max%d' % freq] = np.max(magFFT[freq:freq +
7500])
for p in [10]:
feature_dict[f'num_peaks_{p}'] = feature_calculators.number_peaks(
x, 10)
feature_dict['cid_ce'] = feature_calculators.cid_ce(x, normalize=True)
for w in [5]:
feature_dict[
f'autocorrelation_{w}'] = feature_calculators.autocorrelation(
x, w)
return feature_dict
def preproc(d):
df = pd.DataFrame(d)
x_autocorr = df.apply(lambda x: x.autocorr(lag=5), axis=1)
x_mean = df.apply(lambda x: np.mean(x), axis=1)
x_max = df.apply(lambda x: np.max(x), axis=1)
x_c3 = df.apply(lambda x: tsf_calc.c3(x, 5), axis=1)
x_cid = standardize(df.apply(lambda x: tsf_calc.cid_ce(x, False), axis=1))
x_sym = df.apply(lambda x: 0
if tsf_calc.symmetry_looking(x, [{
'r': 0.0106
}])[0][1] else 1,
axis=1) # all observations that are strongly asymmetric
return pd.concat([x_autocorr, x_mean, x_max, x_c3, x_cid, x_sym], axis=1)
def CIDCELag3(fragment):
return fc.cid_ce(fragment,3)
def calculate_complexity_estimation(traffic):
return feature_calculators.cid_ce(traffic, normalize=True)
def function(x):
return cid_ce(x, normalize=self.normalize)
def features(self, x, prefix):
feature_dict = dict()
# create features here
# numpy
feature_dict[prefix + '_' + 'mean'] = np.mean(x)
feature_dict[prefix + '_' + 'max'] = np.max(x)
feature_dict[prefix + '_' + 'min'] = np.min(x)
feature_dict[prefix + '_' + 'std'] = np.std(x)
feature_dict[prefix + '_' + 'var'] = np.var(x)
feature_dict[prefix + '_' + 'ptp'] = np.ptp(x)
feature_dict[prefix + '_' + 'percentile_10'] = np.percentile(x, 10)
feature_dict[prefix + '_' + 'percentile_20'] = np.percentile(x, 20)
feature_dict[prefix + '_' + 'percentile_30'] = np.percentile(x, 30)
feature_dict[prefix + '_' + 'percentile_40'] = np.percentile(x, 40)
feature_dict[prefix + '_' + 'percentile_50'] = np.percentile(x, 50)
feature_dict[prefix + '_' + 'percentile_60'] = np.percentile(x, 60)
feature_dict[prefix + '_' + 'percentile_70'] = np.percentile(x, 70)
feature_dict[prefix + '_' + 'percentile_80'] = np.percentile(x, 80)
feature_dict[prefix + '_' + 'percentile_90'] = np.percentile(x, 90)
# scipy
feature_dict[prefix + '_' + 'skew'] = sp.stats.skew(x)
feature_dict[prefix + '_' + 'kurtosis'] = sp.stats.kurtosis(x)
feature_dict[prefix + '_' + 'kstat_1'] = sp.stats.kstat(x, 1)
feature_dict[prefix + '_' + 'kstat_2'] = sp.stats.kstat(x, 2)
feature_dict[prefix + '_' + 'kstat_3'] = sp.stats.kstat(x, 3)
feature_dict[prefix + '_' + 'kstat_4'] = sp.stats.kstat(x, 4)
feature_dict[prefix + '_' + 'moment_1'] = sp.stats.moment(x, 1)
feature_dict[prefix + '_' + 'moment_2'] = sp.stats.moment(x, 2)
feature_dict[prefix + '_' + 'moment_3'] = sp.stats.moment(x, 3)
feature_dict[prefix + '_' + 'moment_4'] = sp.stats.moment(x, 4)
# tsfresh
feature_dict[prefix + '_' +
'abs_energy'] = feature_calculators.abs_energy(x)
feature_dict[
prefix + '_' +
'abs_sum_of_changes'] = feature_calculators.absolute_sum_of_changes(
x)
feature_dict[
prefix + '_' +
'count_above_mean'] = feature_calculators.count_above_mean(x)
feature_dict[
prefix + '_' +
'count_below_mean'] = feature_calculators.count_below_mean(x)
feature_dict[prefix + '_' +
'mean_abs_change'] = feature_calculators.mean_abs_change(
x)
feature_dict[prefix + '_' +
'mean_change'] = feature_calculators.mean_change(x)
feature_dict[
prefix + '_' +
'var_larger_than_std_dev'] = feature_calculators.variance_larger_than_standard_deviation(
x)
feature_dict[prefix + '_' +
'range_minf_m4000'] = feature_calculators.range_count(
x, -np.inf, -4000)
feature_dict[prefix + '_' +
'range_m4000_m3000'] = feature_calculators.range_count(
x, -4000, -3000)
feature_dict[prefix + '_' +
'range_m3000_m2000'] = feature_calculators.range_count(
x, -3000, -2000)
feature_dict[prefix + '_' +
'range_m2000_m1000'] = feature_calculators.range_count(
x, -2000, -1000)
feature_dict[prefix + '_' +
'range_m1000_0'] = feature_calculators.range_count(
x, -1000, 0)
feature_dict[prefix + '_' +
'range_0_p1000'] = feature_calculators.range_count(
x, 0, 1000)
feature_dict[prefix + '_' +
'range_p1000_p2000'] = feature_calculators.range_count(
x, 1000, 2000)
feature_dict[prefix + '_' +
'range_p2000_p3000'] = feature_calculators.range_count(
x, 2000, 3000)
feature_dict[prefix + '_' +
'range_p3000_p4000'] = feature_calculators.range_count(
x, 3000, 4000)
feature_dict[prefix + '_' +
'range_p4000_pinf'] = feature_calculators.range_count(
x, 4000, np.inf)
feature_dict[
prefix + '_' +
'ratio_unique_values'] = feature_calculators.ratio_value_number_to_time_series_length(
x)
feature_dict[
prefix + '_' +
'first_loc_min'] = feature_calculators.first_location_of_minimum(x)
feature_dict[
prefix + '_' +
'first_loc_max'] = feature_calculators.first_location_of_maximum(x)
feature_dict[
prefix + '_' +
'last_loc_min'] = feature_calculators.last_location_of_minimum(x)
feature_dict[
prefix + '_' +
'last_loc_max'] = feature_calculators.last_location_of_maximum(x)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_10'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 10)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_100'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 100)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_1000'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 1000)
feature_dict[
prefix + '_' +
'autocorrelation_1'] = feature_calculators.autocorrelation(x, 1)
feature_dict[
prefix + '_' +
'autocorrelation_2'] = feature_calculators.autocorrelation(x, 2)
feature_dict[
prefix + '_' +
'autocorrelation_3'] = feature_calculators.autocorrelation(x, 3)
feature_dict[
prefix + '_' +
'autocorrelation_4'] = feature_calculators.autocorrelation(x, 4)
feature_dict[
prefix + '_' +
'autocorrelation_5'] = feature_calculators.autocorrelation(x, 5)
feature_dict[
prefix + '_' +
'autocorrelation_6'] = feature_calculators.autocorrelation(x, 6)
feature_dict[
prefix + '_' +
'autocorrelation_7'] = feature_calculators.autocorrelation(x, 7)
feature_dict[
prefix + '_' +
'autocorrelation_8'] = feature_calculators.autocorrelation(x, 8)
feature_dict[
prefix + '_' +
'autocorrelation_9'] = feature_calculators.autocorrelation(x, 9)
feature_dict[
prefix + '_' +
'autocorrelation_10'] = feature_calculators.autocorrelation(x, 10)
feature_dict[
prefix + '_' +
'autocorrelation_50'] = feature_calculators.autocorrelation(x, 50)
feature_dict[
prefix + '_' +
'autocorrelation_100'] = feature_calculators.autocorrelation(
x, 100)
feature_dict[
prefix + '_' +
'autocorrelation_1000'] = feature_calculators.autocorrelation(
x, 1000)
feature_dict[prefix + '_' + 'c3_1'] = feature_calculators.c3(x, 1)
feature_dict[prefix + '_' + 'c3_2'] = feature_calculators.c3(x, 2)
feature_dict[prefix + '_' + 'c3_3'] = feature_calculators.c3(x, 3)
feature_dict[prefix + '_' + 'c3_4'] = feature_calculators.c3(x, 4)
feature_dict[prefix + '_' + 'c3_5'] = feature_calculators.c3(x, 5)
feature_dict[prefix + '_' + 'c3_10'] = feature_calculators.c3(x, 10)
feature_dict[prefix + '_' + 'c3_100'] = feature_calculators.c3(x, 100)
for c in range(1, 34):
feature_dict[prefix + '_' + 'fft_{0}_real'.format(c)] = list(
feature_calculators.fft_coefficient(x, [{
'coeff': c,
'attr': 'real'
}]))[0][1]
feature_dict[prefix + '_' + 'fft_{0}_imag'.format(c)] = list(
feature_calculators.fft_coefficient(x, [{
'coeff': c,
'attr': 'imag'
}]))[0][1]
feature_dict[prefix + '_' + 'fft_{0}_ang'.format(c)] = list(
feature_calculators.fft_coefficient(x, [{
'coeff': c,
'attr': 'angle'
}]))[0][1]
feature_dict[
prefix + '_' +
'long_strk_above_mean'] = feature_calculators.longest_strike_above_mean(
x)
feature_dict[
prefix + '_' +
'long_strk_below_mean'] = feature_calculators.longest_strike_below_mean(
x)
feature_dict[prefix + '_' + 'cid_ce_0'] = feature_calculators.cid_ce(
x, 0)
feature_dict[prefix + '_' + 'cid_ce_1'] = feature_calculators.cid_ce(
x, 1)
feature_dict[prefix + '_' +
'binned_entropy_5'] = feature_calculators.binned_entropy(
x, 5)
feature_dict[prefix + '_' +
'binned_entropy_10'] = feature_calculators.binned_entropy(
x, 10)
feature_dict[prefix + '_' +
'binned_entropy_20'] = feature_calculators.binned_entropy(
x, 20)
feature_dict[prefix + '_' +
'binned_entropy_50'] = feature_calculators.binned_entropy(
x, 50)
feature_dict[prefix + '_' +
'binned_entropy_80'] = feature_calculators.binned_entropy(
x, 80)
feature_dict[
prefix + '_' +
'binned_entropy_100'] = feature_calculators.binned_entropy(x, 100)
feature_dict[prefix + '_' +
'num_crossing_0'] = feature_calculators.number_crossing_m(
x, 0)
feature_dict[prefix + '_' +
'num_peaks_1'] = feature_calculators.number_peaks(x, 1)
feature_dict[prefix + '_' +
'num_peaks_3'] = feature_calculators.number_peaks(x, 3)
feature_dict[prefix + '_' +
'num_peaks_5'] = feature_calculators.number_peaks(x, 5)
feature_dict[prefix + '_' +
'num_peaks_10'] = feature_calculators.number_peaks(x, 10)
feature_dict[prefix + '_' +
'num_peaks_50'] = feature_calculators.number_peaks(x, 50)
feature_dict[prefix + '_' +
'num_peaks_100'] = feature_calculators.number_peaks(
x, 100)
feature_dict[prefix + '_' +
'num_peaks_500'] = feature_calculators.number_peaks(
x, 500)
feature_dict[prefix + '_' + 'spkt_welch_density_1'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 1
}]))[0][1]
feature_dict[prefix + '_' + 'spkt_welch_density_2'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 2
}]))[0][1]
feature_dict[prefix + '_' + 'spkt_welch_density_5'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 5
}]))[0][1]
feature_dict[prefix + '_' + 'spkt_welch_density_8'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 8
}]))[0][1]
feature_dict[prefix + '_' + 'spkt_welch_density_10'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 10
}]))[0][1]
feature_dict[prefix + '_' + 'spkt_welch_density_50'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 50
}]))[0][1]
feature_dict[prefix + '_' + 'spkt_welch_density_100'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 100
}]))[0][1]
feature_dict[
prefix + '_' +
'time_rev_asym_stat_1'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 1)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_2'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 2)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_3'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 3)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_4'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 4)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_10'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 10)
feature_dict[
prefix + '_' +
'time_rev_asym_stat_100'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 100)
for r in range(20):
feature_dict[prefix + '_' + 'symmetry_looking_' +
str(r)] = feature_calculators.symmetry_looking(
x, [{
'r': r * 0.05
}])[0][1]
for r in range(1, 20):
feature_dict[
prefix + '_' + 'large_standard_deviation_' +
str(r)] = feature_calculators.large_standard_deviation(
x, r * 0.05)
for r in range(1, 10):
feature_dict[prefix + '_' + 'quantile_' +
str(r)] = feature_calculators.quantile(x, r * 0.1)
for r in ['mean', 'median', 'var']:
feature_dict[prefix + '_' + 'agg_autocorr_' +
r] = feature_calculators.agg_autocorrelation(
x, [{
'f_agg': r,
'maxlag': 40
}])[0][-1]
#for r in range(1, 6):
# feature_dict[prefix+'_'+'number_cwt_peaks_'+str(r)] = feature_calculators.number_cwt_peaks(x, r)
for r in range(1, 10):
feature_dict[prefix + '_' + 'index_mass_quantile_' +
str(r)] = feature_calculators.index_mass_quantile(
x, [{
'q': r
}])[0][1]
#for ql in [0., .2, .4, .6, .8]:
# for qh in [.2, .4, .6, .8, 1.]:
# if ql < qh:
# for b in [False, True]:
# for f in ["mean", "var"]:
# feature_dict[prefix+'_'+'change_quantiles_'+str(ql)+'_'+str(qh)+'_'+str(b)+'_'+str(f)] = feature_calculators.change_quantiles(x, ql, qh, b, f)
#for r in [.1, .3, .5, .7, .9]:
# feature_dict[prefix+'_'+'approximate_entropy_'+str(r)] = feature_calculators.approximate_entropy(x, 2, r)
feature_dict[
prefix + '_' +
'max_langevin_fixed_point'] = feature_calculators.max_langevin_fixed_point(
x, 3, 30)
for r in ['pvalue', 'rvalue', 'intercept', 'slope', 'stderr']:
feature_dict[prefix + '_' + 'linear_trend_' +
str(r)] = feature_calculators.linear_trend(
x, [{
'attr': r
}])[0][1]
for r in ['pvalue', 'teststat', 'usedlag']:
feature_dict[prefix + '_' + 'augmented_dickey_fuller_' +
r] = feature_calculators.augmented_dickey_fuller(
x, [{
'attr': r
}])[0][1]
for r in [0.5, 1, 1.5, 2, 2.5, 3, 5, 6, 7, 10]:
feature_dict[prefix + '_' + 'ratio_beyond_r_sigma_' +
str(r)] = feature_calculators.ratio_beyond_r_sigma(
x, r)
#for attr in ["pvalue", "rvalue", "intercept", "slope", "stderr"]:
# feature_dict[prefix+'_'+'linear_trend_timewise_'+attr] = feature_calculators.linear_trend_timewise(x, [{'attr': attr}])[0][1]
#for attr in ["rvalue", "intercept", "slope", "stderr"]:
# for i in [5, 10, 50]:
# for f in ["max", "min", "mean", "var"]:
# feature_dict[prefix+'_'+'agg_linear_trend_'+attr+'_'+str(i)+'_'+f] = feature_calculators.agg_linear_trend(x, [{'attr': attr, 'chunk_len': i, 'f_agg': f}])[0][-1]
#for width in [2, 5, 10, 20]:
# for coeff in range(15):
# for w in [2, 5, 10, 20]:
# feature_dict[prefix+'_'+'cwt_coefficients_'+str(width)+'_'+str(coeff)+'_'+str(w)] = list(feature_calculators.cwt_coefficients(x, [{'widths': width, 'coeff': coeff, 'w': w}]))[0][1]
#for r in range(10):
# feature_dict[prefix+'_'+'partial_autocorr_'+str(r)] = feature_calculators.partial_autocorrelation(x, [{'lag': r}])[0][1]
# "ar_coefficient": [{"coeff": coeff, "k": k} for coeff in range(5) for k in [10]],
# "fft_coefficient": [{"coeff": k, "attr": a} for a, k in product(["real", "imag", "abs", "angle"], range(100))],
# "fft_aggregated": [{"aggtype": s} for s in ["centroid", "variance", "skew", "kurtosis"]],
# "value_count": [{"value": value} for value in [0, 1, -1]],
# "range_count": [{"min": -1, "max": 1}, {"min": 1e12, "max": 0}, {"min": 0, "max": 1e12}],
# "friedrich_coefficients": (lambda m: [{"coeff": coeff, "m": m, "r": 30} for coeff in range(m + 1)])(3),
# "energy_ratio_by_chunks": [{"num_segments": 10, "segment_focus": i} for i in range(10)],
return feature_dict
def generate_time_series_feats(x_dataset, dataset_name="raw", test=False):
make_dir_if_not_exists(os.path.join(FEATURES_PATH, 'tsfeats'))
time_length = x_dataset.shape[1]
features_function_dict = {
"mean":
mean,
"median":
median,
"length":
length,
"minimum":
minimum,
"maximum":
maximum,
"variance":
variance,
"skewness":
skewness,
"kurtosis":
kurtosis,
"sum_values":
sum_values,
"abs_energy":
abs_energy,
"mean_change":
mean_change,
"mean_abs_change":
mean_abs_change,
"count_below_mean":
count_below_mean,
"count_above_mean":
count_above_mean,
"has_duplicate_min":
has_duplicate_min,
"has_duplicate_max":
has_duplicate_max,
"standard_deviation":
standard_deviation,
"absolute_sum_of_changes":
absolute_sum_of_changes,
"last_location_of_minimum":
last_location_of_minimum,
"last_location_of_maximum":
last_location_of_maximum,
"first_location_of_maximum":
first_location_of_maximum,
"longest_strike_below_mean":
longest_strike_below_mean,
"longest_strike_above_mean":
longest_strike_above_mean,
"sum_of_reoccurring_values":
sum_of_reoccurring_values,
"first_location_of_minimum":
first_location_of_minimum,
"sum_of_reoccurring_data_points":
sum_of_reoccurring_data_points,
"variance_larger_than_standard_deviation":
variance_larger_than_standard_deviation,
"ratio_value_number_to_time_series_length":
ratio_value_number_to_time_series_length,
"percentage_of_reoccurring_values_to_all_values":
percentage_of_reoccurring_values_to_all_values,
"binned_entropy_max300":
lambda x: binned_entropy(x, 300),
"binned_entropy_max400":
lambda x: binned_entropy(x, 400),
"cid_ce_true":
lambda x: cid_ce(x, True),
"cid_ce_false":
lambda x: cid_ce(x, False),
"percentage_of_reoccurring_datapoints_to_all_datapoints":
percentage_of_reoccurring_datapoints_to_all_datapoints
}
for feature_name, function_call in features_function_dict.iteritems():
print "{:.<70s}".format("- Processing feature: %s" % feature_name),
feature_name = 'tsfeats/%s_%s' % (dataset_name, feature_name)
if not features_exists(feature_name, test):
feats = x_dataset.apply(function_call, axis=1, raw=True).values
save_features(feats, feature_name, test)
print("Done")
else:
print("Already generated")
ar_param_k100 = [{"coeff": i, "k": 100} for i in range(100 + 1)]
ar_param_k500 = [{"coeff": i, "k": 500} for i in range(500 + 1)]
agg50_mean_linear_trend = [{
"attr": val,
"chunk_len": 50,
"f_agg": "mean"
} for val in ("pvalue", "rvalue", "intercept", "slope", "stderr")]
aug_dickey_fuler_params = [{
"attr": "teststat"
}, {
"attr": "pvalue"
}, {
"attr": "usedlag"
}]
energy_ratio_num10_focus5 = [{"num_segments": 10, "segment_focus": 5}]
fft_aggr_spectrum = [{
"aggtype": "centroid"
}, {
"aggtype": "variance"
}, {
"aggtype": "skew"
}, {
"aggtype": "kurtosis"
}]
fft_coefficient_real = [{
"coeff": i,
"attr": "real"
} for i in range((time_length + 1) // 2)]
fft_coefficient_imag = [{
"coeff": i,
"attr": "imag"
} for i in range((time_length + 1) // 2)]
fft_coefficient_abs = [{
"coeff": i,
"attr": "abs"
} for i in range((time_length + 1) // 2)]
fft_coefficient_angle = [{
"coeff": i,
"attr": "angle"
} for i in range((time_length + 1) // 2)]
linear_trend_params = [{
"attr": val
} for val in ("pvalue", "rvalue", "intercept", "slope", "stderr")]
other_feats_dict = {
"ar_coeff100":
lambda x: dict(ar_coefficient(x, ar_param_k100)),
"ar_coeff500":
lambda x: dict(ar_coefficient(x, ar_param_k500)),
"agg50_mean_lin_trend":
lambda x: dict(agg_linear_trend(x, agg50_mean_linear_trend)),
"aug_dickey_fuler":
lambda x: dict(augmented_dickey_fuller(x, aug_dickey_fuler_params)),
"energy_ratio_num10_focus5":
lambda x: dict(energy_ratio_by_chunks(x, energy_ratio_num10_focus5)),
"fft_aggr_spectrum":
lambda x: dict(fft_aggregated(x, fft_aggr_spectrum)),
"fft_coeff_real":
lambda x: dict(fft_coefficient(x, fft_coefficient_real)),
"fft_coeff_imag":
lambda x: dict(fft_coefficient(x, fft_coefficient_imag)),
"fft_coeff_abs":
lambda x: dict(fft_coefficient(x, fft_coefficient_abs)),
"fft_coeff_angle":
lambda x: dict(fft_coefficient(x, fft_coefficient_angle)),
"linear_trend":
lambda x: dict(linear_trend(x, linear_trend_params)),
}
for feature_name, function_call in other_feats_dict.iteritems():
print "{:.<70s}".format("- Processing features: %s" % feature_name),
feature_name = 'tsfeats/%s_%s' % (dataset_name, feature_name)
if not features_exists(feature_name, test):
feats_dict = x_dataset.apply(function_call, axis=1,
raw=True).values.tolist()
feats = pd.DataFrame.from_dict(feats_dict)
save_features(feats.values, feature_name, test)
print("Done")
else:
print("Already generated")
# Auto-correlations as features
print("- Processing Auto-correlation features...")
corr_dataset = x_dataset.apply(autocorrelation_all, axis=1, raw=True)
save_features(corr_dataset.values,
'%s_auto_correlation_all' % dataset_name, test)
print("- Processing ARIMA(5,5,1) Features...")
arima_features = parallelize_row(x_dataset.values,
generate_arima_feats,
n_jobs=2)
assert arima_features.shape[0] == x_dataset.shape[0] # Assert the axis
save_features(arima_features, '%s_arima_5_5_1' % dataset_name, test)
def CIDCELag11(fragment):
return fc.cid_ce(fragment,11)
def transform_pack3(df):
""" augment X from tsfresh features"""
x = df.values
output = {}
output['kstat_1'] = stats.kstat(x, 1)
output['kstat_2'] = stats.kstat(x, 2)
output['kstat_3'] = stats.kstat(x, 3)
output['kstat_4'] = stats.kstat(x, 4)
output['abs_energy'] = feature_calculators.abs_energy(x)
output['abs_sum_of_changes'] = feature_calculators.absolute_sum_of_changes(
x)
output['count_above_mean'] = feature_calculators.count_above_mean(x)
output['count_below_mean'] = feature_calculators.count_below_mean(x)
output['range_minf_m4000'] = feature_calculators.range_count(
x, -np.inf, -4000)
output['range_m4000_m3000'] = feature_calculators.range_count(
x, -4000, -3000)
output['range_m3000_m2000'] = feature_calculators.range_count(
x, -3000, -2000)
output['range_m2000_m1000'] = feature_calculators.range_count(
x, -2000, -1000)
output['range_m1000_0'] = feature_calculators.range_count(x, -1000, 0)
output['range_0_p1000'] = feature_calculators.range_count(x, 0, 1000)
output['range_p1000_p2000'] = feature_calculators.range_count(
x, 1000, 2000)
output['range_p2000_p3000'] = feature_calculators.range_count(
x, 2000, 3000)
output['range_p3000_p4000'] = feature_calculators.range_count(
x, 3000, 4000)
output['range_p4000_pinf'] = feature_calculators.range_count(
x, 4000, np.inf)
output[
'ratio_unique_values'] = feature_calculators.ratio_value_number_to_time_series_length(
x)
output['first_loc_min'] = feature_calculators.first_location_of_minimum(x)
output['first_loc_max'] = feature_calculators.first_location_of_maximum(x)
output['last_loc_min'] = feature_calculators.last_location_of_minimum(x)
output['last_loc_max'] = feature_calculators.last_location_of_maximum(x)
output[
'time_rev_asym_stat_10'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 10)
output[
'time_rev_asym_stat_100'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 100)
output[
'time_rev_asym_stat_1000'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 1000)
output['autocorrelation_10'] = feature_calculators.autocorrelation(x, 10)
output['autocorrelation_100'] = feature_calculators.autocorrelation(x, 100)
output['autocorrelation_1000'] = feature_calculators.autocorrelation(
x, 1000)
output['autocorrelation_5000'] = feature_calculators.autocorrelation(
x, 5000)
output['c3_5'] = feature_calculators.c3(x, 5)
output['c3_10'] = feature_calculators.c3(x, 10)
output['c3_100'] = feature_calculators.c3(x, 100)
output[
'long_strk_above_mean'] = feature_calculators.longest_strike_above_mean(
x)
output[
'long_strk_below_mean'] = feature_calculators.longest_strike_below_mean(
x)
output['cid_ce_0'] = feature_calculators.cid_ce(x, 0)
output['cid_ce_1'] = feature_calculators.cid_ce(x, 1)
output['binned_entropy_10'] = feature_calculators.binned_entropy(x, 10)
output['binned_entropy_50'] = feature_calculators.binned_entropy(x, 50)
output['binned_entropy_80'] = feature_calculators.binned_entropy(x, 80)
output['binned_entropy_100'] = feature_calculators.binned_entropy(x, 100)
tmp = np.abs(x)
output['num_crossing_0'] = feature_calculators.number_crossing_m(tmp, 0)
output['num_crossing_10'] = feature_calculators.number_crossing_m(tmp, 10)
output['num_crossing_100'] = feature_calculators.number_crossing_m(
tmp, 100)
output['num_peaks_10'] = feature_calculators.number_peaks(tmp, 10)
output['num_peaks_50'] = feature_calculators.number_peaks(tmp, 50)
output['num_peaks_100'] = feature_calculators.number_peaks(tmp, 100)
output['num_peaks_500'] = feature_calculators.number_peaks(tmp, 500)
output['spkt_welch_density_1'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 1
}]))[0][1]
output['spkt_welch_density_10'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 10
}]))[0][1]
output['spkt_welch_density_50'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 50
}]))[0][1]
output['spkt_welch_density_100'] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': 100
}]))[0][1]
output[
'time_rev_asym_stat_1'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 1)
output[
'time_rev_asym_stat_10'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 10)
output[
'time_rev_asym_stat_100'] = feature_calculators.time_reversal_asymmetry_statistic(
x, 100)
return output
def features(self, x, y, seg_id, denoise=False):
if (denoise == True):
x_hp = high_pass_filter(x, low_cutoff=10000, sample_rate=4000000)
x = denoise_signal(x_hp, wavelet='haar', level=1)
feature_dict = dict()
feature_dict['target'] = y
feature_dict['seg_id'] = seg_id
# create features here
# lists with parameters to iterate over them
percentiles = [
1, 5, 10, 20, 25, 30, 40, 50, 60, 70, 75, 80, 90, 95, 99
]
hann_windows = [50, 150, 1500, 15000]
spans = [300, 3000, 30000, 50000]
windows = [10, 50, 100, 500, 1000, 10000]
borders = list(range(-4000, 4001, 1000))
peaks = [10, 20, 50, 100]
coefs = [1, 5, 10, 50, 100]
lags = [10, 100, 1000, 10000]
autocorr_lags = [5, 10, 50, 100, 500, 1000, 5000, 10000]
# basic stats
feature_dict['mean'] = x.mean()
feature_dict['std'] = x.std()
feature_dict['max'] = x.max()
feature_dict['min'] = x.min()
# basic stats on absolute values
feature_dict['mean_change_abs'] = np.mean(np.diff(x))
feature_dict['abs_max'] = np.abs(x).max()
feature_dict['abs_mean'] = np.abs(x).mean()
feature_dict['abs_std'] = np.abs(x).std()
# geometric and harminic means
feature_dict['hmean'] = stats.hmean(np.abs(x[np.nonzero(x)[0]]))
feature_dict['gmean'] = stats.gmean(np.abs(x[np.nonzero(x)[0]]))
# k-statistic and moments
for i in range(1, 5):
feature_dict['kstat_{}'.format(i)] = stats.kstat(x, i)
feature_dict['moment_{}'.format(i)] = stats.moment(x, i)
for i in [1, 2]:
feature_dict['kstatvar_{}'.format(i)] = stats.kstatvar(x, i)
# aggregations on various slices of data
for agg_type, slice_length, direction in product(
['std', 'min', 'max', 'mean'], [1000, 10000, 50000],
['first', 'last']):
if direction == 'first':
feature_dict['{}_{}_{}'.format(
agg_type, direction,
slice_length)] = x[:slice_length].agg(agg_type)
elif direction == 'last':
feature_dict['{}_{}_{}'.format(
agg_type, direction,
slice_length)] = x[-slice_length:].agg(agg_type)
feature_dict['max_to_min'] = x.max() / np.abs(x.min())
feature_dict['max_to_min_diff'] = x.max() - np.abs(x.min())
feature_dict['count_big'] = len(x[np.abs(x) > 500])
feature_dict['sum'] = x.sum()
feature_dict['mean_change_rate'] = calc_change_rate(x)
# calc_change_rate on slices of data
for slice_length, direction in product([1000, 10000, 50000],
['first', 'last']):
if direction == 'first':
feature_dict['mean_change_rate_{}_{}'.format(
direction,
slice_length)] = calc_change_rate(x[:slice_length])
elif direction == 'last':
feature_dict['mean_change_rate_{}_{}'.format(
direction,
slice_length)] = calc_change_rate(x[-slice_length:])
# percentiles on original and absolute values
for p in percentiles:
feature_dict['percentile_{}'.format(p)] = np.percentile(x, p)
feature_dict['abs_percentile_{}'.format(p)] = np.percentile(
np.abs(x), p)
feature_dict['trend'] = add_trend_feature(x)
feature_dict['abs_trend'] = add_trend_feature(x, abs_values=True)
feature_dict['mad'] = x.mad()
feature_dict['kurt'] = x.kurtosis()
feature_dict['skew'] = x.skew()
feature_dict['med'] = x.median()
feature_dict['Hilbert_mean'] = np.abs(hilbert(x)).mean()
for hw in hann_windows:
feature_dict['Hann_window_mean_{}'.format(hw)] = (
convolve(x, hann(hw), mode='same') / sum(hann(hw))).mean()
feature_dict['classic_sta_lta1_mean'] = classic_sta_lta(x, 500,
10000).mean()
feature_dict['classic_sta_lta2_mean'] = classic_sta_lta(
x, 5000, 100000).mean()
feature_dict['classic_sta_lta3_mean'] = classic_sta_lta(x, 3333,
6666).mean()
feature_dict['classic_sta_lta4_mean'] = classic_sta_lta(
x, 10000, 25000).mean()
feature_dict['classic_sta_lta5_mean'] = classic_sta_lta(x, 50,
1000).mean()
feature_dict['classic_sta_lta6_mean'] = classic_sta_lta(x, 100,
5000).mean()
feature_dict['classic_sta_lta7_mean'] = classic_sta_lta(x, 333,
666).mean()
feature_dict['classic_sta_lta8_mean'] = classic_sta_lta(
x, 4000, 10000).mean()
# exponential rolling statistics
ewma = pd.Series.ewm
for s in spans:
feature_dict['exp_Moving_average_{}_mean'.format(s)] = (ewma(
x, span=s).mean(skipna=True)).mean(skipna=True)
feature_dict['exp_Moving_average_{}_std'.format(s)] = (ewma(
x, span=s).mean(skipna=True)).std(skipna=True)
feature_dict['exp_Moving_std_{}_mean'.format(s)] = (ewma(
x, span=s).std(skipna=True)).mean(skipna=True)
feature_dict['exp_Moving_std_{}_std'.format(s)] = (ewma(
x, span=s).std(skipna=True)).std(skipna=True)
feature_dict['iqr1'] = np.subtract(*np.percentile(x, [95, 5]))
feature_dict['ave10'] = stats.trim_mean(x, 0.1)
for slice_length, threshold in product([50000, 100000, 150000],
[5, 10, 20, 50, 100]):
feature_dict['count_big_{}_threshold_{}'.format(
slice_length,
threshold)] = (np.abs(x[-slice_length:]) > threshold).sum()
feature_dict['count_big_{}_less_threshold_{}'.format(
slice_length,
threshold)] = (np.abs(x[-slice_length:]) < threshold).sum()
# tfresh features take too long to calculate, so I comment them for now
feature_dict['abs_energy'] = feature_calculators.abs_energy(x)
feature_dict[
'abs_sum_of_changes'] = feature_calculators.absolute_sum_of_changes(
x)
feature_dict[
'count_above_mean'] = feature_calculators.count_above_mean(x)
feature_dict[
'count_below_mean'] = feature_calculators.count_below_mean(x)
feature_dict['mean_abs_change'] = feature_calculators.mean_abs_change(
x)
feature_dict['mean_change'] = feature_calculators.mean_change(x)
feature_dict[
'var_larger_than_std_dev'] = feature_calculators.variance_larger_than_standard_deviation(
x)
feature_dict['range_minf_m4000'] = feature_calculators.range_count(
x, -np.inf, -4000)
feature_dict['range_p4000_pinf'] = feature_calculators.range_count(
x, 4000, np.inf)
for i, j in zip(borders, borders[1:]):
feature_dict['range_{}_{}'.format(
i, j)] = feature_calculators.range_count(x, i, j)
feature_dict[
'ratio_unique_values'] = feature_calculators.ratio_value_number_to_time_series_length(
x)
feature_dict[
'first_loc_min'] = feature_calculators.first_location_of_minimum(x)
feature_dict[
'first_loc_max'] = feature_calculators.first_location_of_maximum(x)
feature_dict[
'last_loc_min'] = feature_calculators.last_location_of_minimum(x)
feature_dict[
'last_loc_max'] = feature_calculators.last_location_of_maximum(x)
for lag in lags:
feature_dict['time_rev_asym_stat_{}'.format(
lag)] = feature_calculators.time_reversal_asymmetry_statistic(
x, lag)
for autocorr_lag in autocorr_lags:
feature_dict['autocorrelation_{}'.format(
autocorr_lag)] = feature_calculators.autocorrelation(
x, autocorr_lag)
feature_dict['c3_{}'.format(
autocorr_lag)] = feature_calculators.c3(x, autocorr_lag)
for coeff, attr in product([1, 2, 3, 4, 5], ['real', 'imag', 'angle']):
feature_dict['fft_{}_{}'.format(coeff, attr)] = list(
feature_calculators.fft_coefficient(x, [{
'coeff': coeff,
'attr': attr
}]))[0][1]
feature_dict[
'long_strk_above_mean'] = feature_calculators.longest_strike_above_mean(
x)
feature_dict[
'long_strk_below_mean'] = feature_calculators.longest_strike_below_mean(
x)
feature_dict['cid_ce_0'] = feature_calculators.cid_ce(x, 0)
feature_dict['cid_ce_1'] = feature_calculators.cid_ce(x, 1)
for p in percentiles:
feature_dict['binned_entropy_{}'.format(
p)] = feature_calculators.binned_entropy(x, p)
feature_dict['num_crossing_0'] = feature_calculators.number_crossing_m(
x, 0)
for peak in peaks:
feature_dict['num_peaks_{}'.format(
peaks)] = feature_calculators.number_peaks(x, peak)
for c in coefs:
feature_dict['spkt_welch_density_{}'.format(c)] = list(
feature_calculators.spkt_welch_density(x, [{
'coeff': c
}]))[0][1]
feature_dict['time_rev_asym_stat_{}'.format(
c)] = feature_calculators.time_reversal_asymmetry_statistic(
x, c)
# statistics on rolling windows of various sizes
for w in windows:
x_roll_std = x.rolling(w).std().dropna().values
x_roll_mean = x.rolling(w).mean().dropna().values
feature_dict['ave_roll_std_{}'.format(w)] = x_roll_std.mean()
feature_dict['std_roll_std_{}'.format(w)] = x_roll_std.std()
feature_dict['max_roll_std_{}'.format(w)] = x_roll_std.max()
feature_dict['min_roll_std_{}'.format(w)] = x_roll_std.min()
for p in percentiles:
feature_dict['percentile_roll_std_{}_window_{}'.format(
p, w)] = np.percentile(x_roll_std, p)
feature_dict['av_change_abs_roll_std_{}'.format(w)] = np.mean(
np.diff(x_roll_std))
feature_dict['av_change_rate_roll_std_{}'.format(w)] = np.mean(
np.nonzero((np.diff(x_roll_std) / x_roll_std[:-1]))[0])
feature_dict['abs_max_roll_std_{}'.format(w)] = np.abs(
x_roll_std).max()
feature_dict['ave_roll_mean_{}'.format(w)] = x_roll_mean.mean()
feature_dict['std_roll_mean_{}'.format(w)] = x_roll_mean.std()
feature_dict['max_roll_mean_{}'.format(w)] = x_roll_mean.max()
feature_dict['min_roll_mean_{}'.format(w)] = x_roll_mean.min()
for p in percentiles:
feature_dict['percentile_roll_mean_{}_window_{}'.format(
p, w)] = np.percentile(x_roll_mean, p)
feature_dict['av_change_abs_roll_mean_{}'.format(w)] = np.mean(
np.diff(x_roll_mean))
feature_dict['av_change_rate_roll_mean_{}'.format(w)] = np.mean(
np.nonzero((np.diff(x_roll_mean) / x_roll_mean[:-1]))[0])
feature_dict['abs_max_roll_mean_{}'.format(w)] = np.abs(
x_roll_mean).max()
return feature_dict
def create_features(seg_id, seg, X, st, end):
"""
create features including fft features, statistical features and time series features
:param seg_id: the ID for a sample
:param seg: s signal segment
:param X: train set features before creating these features
:param st: the start index of the signal segment
:param end: the end index of the signal segment
:return: train set features after creating these features
"""
try:
# test set won't create these features because its seg_id is string
X.loc[seg_id, 'seg_id'] = np.int32(seg_id)
X.loc[seg_id, 'seg_start'] = np.int32(st)
X.loc[seg_id, 'seg_end'] = np.int32(end)
except ValueError:
pass
xc = pd.Series(seg['acoustic_data'].values)
xcdm = xc - np.mean(xc)
b, a = des_bw_filter_lp(cutoff=18000)
xcz = sg.lfilter(b, a, xcdm)
zc = np.fft.fft(xcz)
zc = zc[:MAX_FREQ]
# FFT transform values
realFFT = np.real(zc)
imagFFT = np.imag(zc)
freq_bands = [x for x in range(0, MAX_FREQ, FREQ_BAND)]
magFFT = np.sqrt(realFFT ** 2 + imagFFT ** 2)
phzFFT = np.arctan(imagFFT / realFFT)
phzFFT[phzFFT == -np.inf] = -np.pi / 2.0
phzFFT[phzFFT == np.inf] = np.pi / 2.0
phzFFT = np.nan_to_num(phzFFT)
for freq in freq_bands:
X.loc[seg_id, 'FFT_Mag_01q%d' % freq] = np.quantile(magFFT[freq: freq + FREQ_BAND], 0.01)
X.loc[seg_id, 'FFT_Mag_10q%d' % freq] = np.quantile(magFFT[freq: freq + FREQ_BAND], 0.1)
X.loc[seg_id, 'FFT_Mag_90q%d' % freq] = np.quantile(magFFT[freq: freq + FREQ_BAND], 0.9)
X.loc[seg_id, 'FFT_Mag_99q%d' % freq] = np.quantile(magFFT[freq: freq + FREQ_BAND], 0.99)
X.loc[seg_id, 'FFT_Mag_mean%d' % freq] = np.mean(magFFT[freq: freq + FREQ_BAND])
X.loc[seg_id, 'FFT_Mag_std%d' % freq] = np.std(magFFT[freq: freq + FREQ_BAND])
X.loc[seg_id, 'FFT_Mag_max%d' % freq] = np.max(magFFT[freq: freq + FREQ_BAND])
X.loc[seg_id, 'FFT_Phz_mean%d' % freq] = np.mean(phzFFT[freq: freq + FREQ_BAND])
X.loc[seg_id, 'FFT_Phz_std%d' % freq] = np.std(phzFFT[freq: freq + FREQ_BAND])
X.loc[seg_id, 'FFT_Rmean'] = realFFT.mean()
X.loc[seg_id, 'FFT_Rstd'] = realFFT.std()
X.loc[seg_id, 'FFT_Rmax'] = realFFT.max()
X.loc[seg_id, 'FFT_Rmin'] = realFFT.min()
X.loc[seg_id, 'FFT_Imean'] = imagFFT.mean()
X.loc[seg_id, 'FFT_Istd'] = imagFFT.std()
X.loc[seg_id, 'FFT_Imax'] = imagFFT.max()
X.loc[seg_id, 'FFT_Imin'] = imagFFT.min()
X.loc[seg_id, 'FFT_Rmean_first_6000'] = realFFT[:6000].mean()
X.loc[seg_id, 'FFT_Rstd__first_6000'] = realFFT[:6000].std()
X.loc[seg_id, 'FFT_Rmax_first_6000'] = realFFT[:6000].max()
X.loc[seg_id, 'FFT_Rmin_first_6000'] = realFFT[:6000].min()
X.loc[seg_id, 'FFT_Rmean_first_18000'] = realFFT[:18000].mean()
X.loc[seg_id, 'FFT_Rstd_first_18000'] = realFFT[:18000].std()
X.loc[seg_id, 'FFT_Rmax_first_18000'] = realFFT[:18000].max()
X.loc[seg_id, 'FFT_Rmin_first_18000'] = realFFT[:18000].min()
del xcz
del zc
b, a = des_bw_filter_lp(cutoff=2500)
xc0 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=2500, high=5000)
xc1 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=5000, high=7500)
xc2 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=7500, high=10000)
xc3 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=10000, high=12500)
xc4 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=12500, high=15000)
xc5 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=15000, high=17500)
xc6 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_bp(low=17500, high=20000)
xc7 = sg.lfilter(b, a, xcdm)
b, a = des_bw_filter_hp(cutoff=20000)
xc8 = sg.lfilter(b, a, xcdm)
sigs = [xc, pd.Series(xc0), pd.Series(xc1), pd.Series(xc2), pd.Series(xc3),
pd.Series(xc4), pd.Series(xc5), pd.Series(xc6), pd.Series(xc7), pd.Series(xc8)]
for i, sig in enumerate(sigs):
X.loc[seg_id, 'mean_%d' % i] = sig.mean()
X.loc[seg_id, 'std_%d' % i] = sig.std()
X.loc[seg_id, 'max_%d' % i] = sig.max()
X.loc[seg_id, 'min_%d' % i] = sig.min()
X.loc[seg_id, 'mean_change_abs_%d' % i] = np.mean(np.diff(sig))
X.loc[seg_id, 'mean_change_rate_%d' % i] = calc_mean_change_rate(sig)
X.loc[seg_id, 'abs_max_%d' % i] = np.abs(sig).max()
X.loc[seg_id, 'std_first_50000_%d' % i] = sig[:50000].std()
X.loc[seg_id, 'std_last_50000_%d' % i] = sig[-50000:].std()
X.loc[seg_id, 'std_first_10000_%d' % i] = sig[:10000].std()
X.loc[seg_id, 'std_last_10000_%d' % i] = sig[-10000:].std()
X.loc[seg_id, 'avg_first_50000_%d' % i] = sig[:50000].mean()
X.loc[seg_id, 'avg_last_50000_%d' % i] = sig[-50000:].mean()
X.loc[seg_id, 'avg_first_10000_%d' % i] = sig[:10000].mean()
X.loc[seg_id, 'avg_last_10000_%d' % i] = sig[-10000:].mean()
X.loc[seg_id, 'min_first_50000_%d' % i] = sig[:50000].min()
X.loc[seg_id, 'min_last_50000_%d' % i] = sig[-50000:].min()
X.loc[seg_id, 'min_first_10000_%d' % i] = sig[:10000].min()
X.loc[seg_id, 'min_last_10000_%d' % i] = sig[-10000:].min()
X.loc[seg_id, 'max_first_50000_%d' % i] = sig[:50000].max()
X.loc[seg_id, 'max_last_50000_%d' % i] = sig[-50000:].max()
X.loc[seg_id, 'max_first_10000_%d' % i] = sig[:10000].max()
X.loc[seg_id, 'max_last_10000_%d' % i] = sig[-10000:].max()
X.loc[seg_id, 'max_to_min_%d' % i] = sig.max() / np.abs(sig.min())
X.loc[seg_id, 'max_to_min_diff_%d' % i] = sig.max() - np.abs(sig.min())
X.loc[seg_id, 'count_big_%d' % i] = len(sig[np.abs(sig) > 500])
X.loc[seg_id, 'mean_change_rate_first_50000_%d' % i] = calc_mean_change_rate(sig[:50000])
X.loc[seg_id, 'mean_change_rate_last_50000_%d' % i] = calc_mean_change_rate(sig[-50000:])
X.loc[seg_id, 'mean_change_rate_first_10000_%d' % i] = calc_mean_change_rate(sig[:10000])
X.loc[seg_id, 'mean_change_rate_last_10000_%d' % i] = calc_mean_change_rate(sig[-10000:])
X.loc[seg_id, 'q95_%d' % i] = np.quantile(sig, 0.95)
X.loc[seg_id, 'q99_%d' % i] = np.quantile(sig, 0.99)
X.loc[seg_id, 'q05_%d' % i] = np.quantile(sig, 0.05)
X.loc[seg_id, 'q01_%d' % i] = np.quantile(sig, 0.01)
X.loc[seg_id, 'abs_q95_%d' % i] = np.quantile(np.abs(sig), 0.95)
X.loc[seg_id, 'abs_q99_%d' % i] = np.quantile(np.abs(sig), 0.99)
X.loc[seg_id, 'abs_q05_%d' % i] = np.quantile(np.abs(sig), 0.05)
X.loc[seg_id, 'abs_q01_%d' % i] = np.quantile(np.abs(sig), 0.01)
X.loc[seg_id, 'trend_%d' % i] = add_trend_feature(sig)
X.loc[seg_id, 'abs_trend_%d' % i] = add_trend_feature(sig, abs_values=True)
X.loc[seg_id, 'abs_mean_%d' % i] = np.abs(sig).mean()
X.loc[seg_id, 'abs_std_%d' % i] = np.abs(sig).std()
X.loc[seg_id, 'mad_%d' % i] = sig.mad()
X.loc[seg_id, 'kurt_%d' % i] = sig.kurtosis()
X.loc[seg_id, 'skew_%d' % i] = sig.skew()
X.loc[seg_id, 'med_%d' % i] = sig.median()
X.loc[seg_id, 'Hilbert_mean_%d' % i] = np.abs(hilbert(sig)).mean()
X.loc[seg_id, 'Hann_window_mean'] = (convolve(xc, hann(150), mode='same') / sum(hann(150))).mean()
X.loc[seg_id, 'classic_sta_lta1_mean_%d' % i] = classic_sta_lta(sig, 500, 10000).mean()
X.loc[seg_id, 'classic_sta_lta2_mean_%d' % i] = classic_sta_lta(sig, 5000, 100000).mean()
X.loc[seg_id, 'classic_sta_lta3_mean_%d' % i] = classic_sta_lta(sig, 3333, 6666).mean()
X.loc[seg_id, 'classic_sta_lta4_mean_%d' % i] = classic_sta_lta(sig, 10000, 25000).mean()
X.loc[seg_id, 'Moving_average_700_mean_%d' % i] = sig.rolling(window=700).mean().mean(skipna=True)
X.loc[seg_id, 'Moving_average_1500_mean_%d' % i] = sig.rolling(window=1500).mean().mean(skipna=True)
X.loc[seg_id, 'Moving_average_3000_mean_%d' % i] = sig.rolling(window=3000).mean().mean(skipna=True)
X.loc[seg_id, 'Moving_average_6000_mean_%d' % i] = sig.rolling(window=6000).mean().mean(skipna=True)
ewma = pd.Series.ewm
X.loc[seg_id, 'exp_Moving_average_300_mean_%d' % i] = ewma(sig, span=300).mean().mean(skipna=True)
X.loc[seg_id, 'exp_Moving_average_3000_mean_%d' % i] = ewma(sig, span=3000).mean().mean(skipna=True)
X.loc[seg_id, 'exp_Moving_average_30000_mean_%d' % i] = ewma(sig, span=30000).mean().mean(skipna=True)
no_of_std = 3
X.loc[seg_id, 'MA_700MA_std_mean_%d' % i] = sig.rolling(window=700).std().mean()
X.loc[seg_id, 'MA_700MA_BB_high_mean_%d' % i] = (
X.loc[seg_id, 'Moving_average_700_mean_%d' % i] + no_of_std * X.loc[
seg_id, 'MA_700MA_std_mean_%d' % i]).mean()
X.loc[seg_id, 'MA_700MA_BB_low_mean_%d' % i] = (
X.loc[seg_id, 'Moving_average_700_mean_%d' % i] - no_of_std * X.loc[
seg_id, 'MA_700MA_std_mean_%d' % i]).mean()
X.loc[seg_id, 'MA_400MA_std_mean_%d' % i] = sig.rolling(window=400).std().mean()
X.loc[seg_id, 'MA_400MA_BB_high_mean_%d' % i] = (
X.loc[seg_id, 'Moving_average_700_mean_%d' % i] + no_of_std * X.loc[
seg_id, 'MA_400MA_std_mean_%d' % i]).mean()
X.loc[seg_id, 'MA_400MA_BB_low_mean_%d' % i] = (
X.loc[seg_id, 'Moving_average_700_mean_%d' % i] - no_of_std * X.loc[
seg_id, 'MA_400MA_std_mean_%d' % i]).mean()
X.loc[seg_id, 'MA_1000MA_std_mean_%d' % i] = sig.rolling(window=1000).std().mean()
X.loc[seg_id, 'iqr_%d' % i] = np.subtract(*np.percentile(sig, [75, 25]))
X.loc[seg_id, 'q999_%d' % i] = np.quantile(sig, 0.999)
X.loc[seg_id, 'q001_%d' % i] = np.quantile(sig, 0.001)
X.loc[seg_id, 'ave10_%d' % i] = stats.trim_mean(sig, 0.1)
X.loc[seg_id, 'num_peaks_10_%d' % i] = feature_calculators.number_peaks(sig, 10)
X.loc[seg_id, 'cid_ce_1_%d' % i] = feature_calculators.cid_ce(sig, 1) # time series complexity
X.loc[seg_id, 'count_1000_0_%d' % i] = feature_calculators.range_count(sig, -1000, 0)
X.loc[seg_id, 'binned_entropy_5_%d' % i] = feature_calculators.binned_entropy(sig, 5)
X.loc[seg_id, 'binned_entropy_15_%d' % i] = feature_calculators.binned_entropy(sig, 15)
# sliding window is a kind of filter, so this code is out of the cycle of band pass
for windows in [10, 100, 1000]:
x_roll_std = xc.rolling(windows).std().dropna()
x_roll_mean = xc.rolling(windows).mean().dropna()
X.loc[seg_id, 'ave_roll_std_' + str(windows)] = x_roll_std.mean()
X.loc[seg_id, 'std_roll_std_' + str(windows)] = x_roll_std.std()
X.loc[seg_id, 'max_roll_std_' + str(windows)] = x_roll_std.max()
X.loc[seg_id, 'min_roll_std_' + str(windows)] = x_roll_std.min()
X.loc[seg_id, 'q01_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.01)
X.loc[seg_id, 'q05_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.05)
X.loc[seg_id, 'q95_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.95)
X.loc[seg_id, 'q99_roll_std_' + str(windows)] = np.quantile(x_roll_std, 0.99)
X.loc[seg_id, 'av_change_abs_roll_std_' + str(windows)] = np.mean(np.diff(x_roll_std))
X.loc[seg_id, 'av_change_rate_roll_std_' + str(windows)] = calc_mean_change_rate(x_roll_std)
X.loc[seg_id, 'abs_max_roll_std_' + str(windows)] = np.abs(x_roll_std).max()
X.loc[seg_id, 'ave_roll_mean_' + str(windows)] = x_roll_mean.mean()
X.loc[seg_id, 'std_roll_mean_' + str(windows)] = x_roll_mean.std()
X.loc[seg_id, 'max_roll_mean_' + str(windows)] = x_roll_mean.max()
X.loc[seg_id, 'min_roll_mean_' + str(windows)] = x_roll_mean.min()
X.loc[seg_id, 'q01_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.01)
X.loc[seg_id, 'q05_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.05)
X.loc[seg_id, 'q95_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.95)
X.loc[seg_id, 'q99_roll_mean_' + str(windows)] = np.quantile(x_roll_mean, 0.99)
X.loc[seg_id, 'av_change_abs_roll_mean_' + str(windows)] = np.mean(np.diff(x_roll_mean))
X.loc[seg_id, 'av_change_rate_roll_mean_' + str(windows)] = calc_mean_change_rate(x_roll_mean)
X.loc[seg_id, 'abs_max_roll_mean_' + str(windows)] = np.abs(x_roll_mean).max()
return X
def CIDCELag5(fragment):
return fc.cid_ce(fragment,5)
def CIDCELag7(fragment):
return fc.cid_ce(fragment,7)
# 3. `last_location_of_maximum`: locates the last occurrence of the maximum value in the time series
# 4. `skewness`: the Fisher-Pearson skewness of the time series
# 5. `sample_entropy`: the sample entropy of the time series
#
# All of these will be calculated for each operational setting and each sensor measurement giving us 5 new columns for each of the 29 original features (145 total features).
# In[ ]:
from tsfresh.feature_extraction.feature_calculators import (
cid_ce, number_peaks, last_location_of_maximum, skewness, sample_entropy)
# To avoid the issue of passing multiple functions to `agg` with the same name `lambda`, we have to create lambda functions and then give them custom names. `cid_ce` and `number_peaks` both have required arguments but the other functions only need a time-series.
# In[ ]:
cid_ce_func = lambda x: cid_ce(x, normalize=False)
cid_ce_func.__name__ = 'cid_ce'
n_peaks = lambda x: number_peaks(x, n=5)
n_peaks.__name__ = 'number_peaks'
# Apply the five operations
ts_values = train_obs.drop(
columns=['time_in_cycles']).groupby('engine_no').agg([
cid_ce_func, n_peaks, last_location_of_maximum, skewness,
sample_entropy
])
ts_values.head()
# Below we rename the columns.
def complexity(x):
return fc.cid_ce(x, True)
def features(self, x, y, seg_id):
feature_dict = dict()
feature_dict['target'] = y
feature_dict['seg_id'] = seg_id
# create features here
# numpy
feature_dict['mean'] = np.mean(x)
feature_dict['max'] = np.max(x)
feature_dict['min'] = np.min(x)
feature_dict['std'] = np.std(x)
feature_dict['var'] = np.var(x)
feature_dict['ptp'] = np.ptp(x)
feature_dict['percentile_10'] = np.percentile(x, 10)
feature_dict['percentile_20'] = np.percentile(x, 20)
feature_dict['percentile_30'] = np.percentile(x, 30)
feature_dict['percentile_40'] = np.percentile(x, 40)
feature_dict['percentile_50'] = np.percentile(x, 50)
feature_dict['percentile_60'] = np.percentile(x, 60)
feature_dict['percentile_70'] = np.percentile(x, 70)
feature_dict['percentile_80'] = np.percentile(x, 80)
feature_dict['percentile_90'] = np.percentile(x, 90)
# scipy
feature_dict['skew'] = sp.stats.skew(x)
feature_dict['kurtosis'] = sp.stats.kurtosis(x)
feature_dict['kstat_1'] = sp.stats.kstat(x, 1)
feature_dict['kstat_2'] = sp.stats.kstat(x, 2)
feature_dict['kstat_3'] = sp.stats.kstat(x, 3)
feature_dict['kstat_4'] = sp.stats.kstat(x, 4)
feature_dict['moment_1'] = sp.stats.moment(x, 1)
feature_dict['moment_2'] = sp.stats.moment(x, 2)
feature_dict['moment_3'] = sp.stats.moment(x, 3)
feature_dict['moment_4'] = sp.stats.moment(x, 4)
feature_dict['abs_energy'] = feature_calculators.abs_energy(x)
feature_dict['abs_sum_of_changes'] = feature_calculators.absolute_sum_of_changes(x)
feature_dict['count_above_mean'] = feature_calculators.count_above_mean(x)
feature_dict['count_below_mean'] = feature_calculators.count_below_mean(x)
feature_dict['mean_abs_change'] = feature_calculators.mean_abs_change(x)
feature_dict['mean_change'] = feature_calculators.mean_change(x)
feature_dict['var_larger_than_std_dev'] = feature_calculators.variance_larger_than_standard_deviation(x)
feature_dict['range_minf_m4000'] = feature_calculators.range_count(x, -np.inf, -4000)
feature_dict['range_m4000_m3000'] = feature_calculators.range_count(x, -4000, -3000)
feature_dict['range_m3000_m2000'] = feature_calculators.range_count(x, -3000, -2000)
feature_dict['range_m2000_m1000'] = feature_calculators.range_count(x, -2000, -1000)
feature_dict['range_m1000_0'] = feature_calculators.range_count(x, -1000, 0)
feature_dict['range_0_p1000'] = feature_calculators.range_count(x, 0, 1000)
feature_dict['range_p1000_p2000'] = feature_calculators.range_count(x, 1000, 2000)
feature_dict['range_p2000_p3000'] = feature_calculators.range_count(x, 2000, 3000)
feature_dict['range_p3000_p4000'] = feature_calculators.range_count(x, 3000, 4000)
feature_dict['range_p4000_pinf'] = feature_calculators.range_count(x, 4000, np.inf)
feature_dict['ratio_unique_values'] = feature_calculators.ratio_value_number_to_time_series_length(x)
feature_dict['first_loc_min'] = feature_calculators.first_location_of_minimum(x)
feature_dict['first_loc_max'] = feature_calculators.first_location_of_maximum(x)
feature_dict['last_loc_min'] = feature_calculators.last_location_of_minimum(x)
feature_dict['last_loc_max'] = feature_calculators.last_location_of_maximum(x)
feature_dict['time_rev_asym_stat_10'] = feature_calculators.time_reversal_asymmetry_statistic(x, 10)
feature_dict['time_rev_asym_stat_100'] = feature_calculators.time_reversal_asymmetry_statistic(x, 100)
feature_dict['time_rev_asym_stat_1000'] = feature_calculators.time_reversal_asymmetry_statistic(x, 1000)
feature_dict['autocorrelation_5'] = feature_calculators.autocorrelation(x, 5)
feature_dict['autocorrelation_10'] = feature_calculators.autocorrelation(x, 10)
feature_dict['autocorrelation_50'] = feature_calculators.autocorrelation(x, 50)
feature_dict['autocorrelation_100'] = feature_calculators.autocorrelation(x, 100)
feature_dict['autocorrelation_1000'] = feature_calculators.autocorrelation(x, 1000)
feature_dict['c3_5'] = feature_calculators.c3(x, 5)
feature_dict['c3_10'] = feature_calculators.c3(x, 10)
feature_dict['c3_100'] = feature_calculators.c3(x, 100)
feature_dict['fft_1_real'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 1, 'attr': 'real'}]))[0][1]
feature_dict['fft_1_imag'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 1, 'attr': 'imag'}]))[0][1]
feature_dict['fft_1_ang'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 1, 'attr': 'angle'}]))[0][1]
feature_dict['fft_2_real'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 2, 'attr': 'real'}]))[0][1]
feature_dict['fft_2_imag'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 2, 'attr': 'imag'}]))[0][1]
feature_dict['fft_2_ang'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 2, 'attr': 'angle'}]))[0][1]
feature_dict['fft_3_real'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 3, 'attr': 'real'}]))[0][1]
feature_dict['fft_3_imag'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 3, 'attr': 'imag'}]))[0][1]
feature_dict['fft_3_ang'] = list(feature_calculators.fft_coefficient(x, [{'coeff': 3, 'attr': 'angle'}]))[0][1]
feature_dict['long_strk_above_mean'] = feature_calculators.longest_strike_above_mean(x)
feature_dict['long_strk_below_mean'] = feature_calculators.longest_strike_below_mean(x)
feature_dict['cid_ce_0'] = feature_calculators.cid_ce(x, 0)
feature_dict['cid_ce_1'] = feature_calculators.cid_ce(x, 1)
feature_dict['binned_entropy_5'] = feature_calculators.binned_entropy(x, 5)
feature_dict['binned_entropy_10'] = feature_calculators.binned_entropy(x, 10)
feature_dict['binned_entropy_20'] = feature_calculators.binned_entropy(x, 20)
feature_dict['binned_entropy_50'] = feature_calculators.binned_entropy(x, 50)
feature_dict['binned_entropy_80'] = feature_calculators.binned_entropy(x, 80)
feature_dict['binned_entropy_100'] = feature_calculators.binned_entropy(x, 100)
feature_dict['num_crossing_0'] = feature_calculators.number_crossing_m(x, 0)
feature_dict['num_peaks_10'] = feature_calculators.number_peaks(x, 10)
feature_dict['num_peaks_50'] = feature_calculators.number_peaks(x, 50)
feature_dict['num_peaks_100'] = feature_calculators.number_peaks(x, 100)
feature_dict['num_peaks_500'] = feature_calculators.number_peaks(x, 500)
feature_dict['spkt_welch_density_1'] = list(feature_calculators.spkt_welch_density(x, [{'coeff': 1}]))[0][1]
feature_dict['spkt_welch_density_10'] = list(feature_calculators.spkt_welch_density(x, [{'coeff': 10}]))[0][1]
feature_dict['spkt_welch_density_50'] = list(feature_calculators.spkt_welch_density(x, [{'coeff': 50}]))[0][1]
feature_dict['spkt_welch_density_100'] = list(feature_calculators.spkt_welch_density(x, [{'coeff': 100}]))[0][1]
feature_dict['time_rev_asym_stat_1'] = feature_calculators.time_reversal_asymmetry_statistic(x, 1)
feature_dict['time_rev_asym_stat_10'] = feature_calculators.time_reversal_asymmetry_statistic(x, 10)
feature_dict['time_rev_asym_stat_100'] = feature_calculators.time_reversal_asymmetry_statistic(x, 100)
return feature_dict
