calculate_timeseries_properties_operator¶
Operator to aggregate column features based on a column
-
class
tasrif.processing_pipeline.kats.calculate_timeseries_properties_operator.CalculateTimeseriesPropertiesOperator(date_feature_name='time', value_column='value', method='kats', **kwargs)¶ This method extracts timeseries features from passed dataframe object
>>> import numpy as np >>> import pandas as pd >>> >>> from tasrif.processing_pipeline.custom import ExtractTimeseriesFeaturesOperator >>> >>> >>> dates = pd.date_range('2016-12-31', '2020-01-08', freq='D').to_series() >>> df = pd.DataFrame() >>> df["Date"] = dates >>> df['Steps'] = np.random.randint(1000,25000, size=len(df)) >>> df['Calories'] = np.random.randint(1800,3000, size=len(df)) >>> df Date Steps Calories 2016-12-31 2016-12-31 14648 2926 2017-01-01 2017-01-01 9320 2190 2017-01-02 2017-01-02 2798 2521 2017-01-03 2017-01-03 11050 2330 2017-01-04 2017-01-04 6536 2172 ... ... ... ... 2020-01-04 2020-01-04 22739 2365 2020-01-05 2020-01-05 4845 1849 2020-01-06 2020-01-06 1143 2420 2020-01-07 2020-01-07 5577 2821 2020-01-08 2020-01-08 10435 1830
>>> operator = ExtractTimeseriesFeaturesOperator(date_feature_name="Date", value_column='Steps') >>> features = operator.process(df)[0] >>> features {'length': 1104, 'mean': 13024.617753623188, 'var': 49311921.9535254, 'entropy': 0.9344372604411008, 'lumpiness': 98331798003649.39, 'stability': 2856199.257016417, 'flat_spots': 1, 'hurst': 0.008804860201927526, 'std1st_der': 4942.8445155096715, 'crossing_points': 559, 'binarize_mean': 0.5, 'unitroot_kpss': 0.040225392515162994, 'heterogeneity': 12.514530846049983, 'histogram_mode': 1008.0, 'linearity': 0.0010899360162310767, 'trend_strength': 0.254325167656127, 'seasonality_strength': 0.3490641758024736, 'spikiness': 757835496.8906168, 'peak': 5, 'trough': 2, 'level_shift_idx': 526, 'level_shift_size': 1180.800000000001, 'y_acf1': 0.04351456010311684, 'y_acf5': 0.004305213630472914, 'diff1y_acf1': -0.4827096027299026, 'diff1y_acf5': 0.23486910790170704, 'diff2y_acf1': -0.657077984628258, 'diff2y_acf5': 0.45760076640469927, 'y_pacf5': 0.004035237116844455, 'diff1y_pacf5': 0.46635744288898123, 'diff2y_pacf5': 1.0429819204003172, 'seas_acf1': 0.0074998374601003975, 'seas_pacf1': 0.010035485775513319, 'firstmin_ac': 3, 'firstzero_ac': 6, 'holt_alpha': 0.21714285714285714, 'holt_beta': 0.09771428571428571, 'hw_alpha': 0.040357142857142855, 'hw_beta': 0.024214285714285716, 'hw_gamma': 0.03427295918367347}
-
__init__(date_feature_name='time', value_column='value', method='kats', **kwargs)¶ Creates a new instance of ExtractTimeseriesFeaturesOperator
- Parameters
date_feature_name – str Name of the datetime column
value_column – str Name of the column that contains values per date
method – str Name of feature extractor method
**kwargs – None or List[str]; list of feature/feature group name(s) key word arguments passed to method’s parameters
-