76 lines
3.9 KiB
Python
76 lines
3.9 KiB
Python
import pandas as pd
|
|
from scipy.stats import entropy
|
|
|
|
|
|
def statsFeatures(eda_data, features, eda_features):
|
|
col_name = "electrodermal_activity"
|
|
if "sumeda" in features:
|
|
eda_features["sumeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].sum()
|
|
if "maxeda" in features:
|
|
eda_features["maxeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].max()
|
|
if "mineda" in features:
|
|
eda_features["mineda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].min()
|
|
if "avgeda" in features:
|
|
eda_features["avgeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].mean()
|
|
if "medianeda" in features:
|
|
eda_features["medianeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].median()
|
|
if "modeeda" in features:
|
|
eda_features["modeeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].agg(lambda x: pd.Series.mode(x)[0])
|
|
if "stdeda" in features:
|
|
eda_features["stdeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].std()
|
|
if "diffmaxmodeeda" in features:
|
|
eda_features["diffmaxmodeeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].max() - \
|
|
eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].agg(lambda x: pd.Series.mode(x)[0])
|
|
if "diffminmodeeda" in features:
|
|
eda_features["diffminmodeeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].agg(lambda x: pd.Series.mode(x)[0]) - \
|
|
eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].min()
|
|
if "entropyeda" in features:
|
|
eda_features["entropyeda"] = eda_data[["local_segment", col_name]].groupby(["local_segment"])[
|
|
col_name].agg(entropy)
|
|
|
|
return eda_features
|
|
|
|
|
|
def extractEDAFeaturesFromIntradayData(eda_intraday_data, features, time_segment, filter_data_by_segment):
|
|
eda_intraday_features = pd.DataFrame(columns=["local_segment"] + features)
|
|
if not eda_intraday_data.empty:
|
|
eda_intraday_data = filter_data_by_segment(eda_intraday_data, time_segment)
|
|
|
|
if not eda_intraday_data.empty:
|
|
eda_intraday_features = pd.DataFrame()
|
|
|
|
# get stats of eda
|
|
eda_intraday_features = statsFeatures(eda_intraday_data, features, eda_intraday_features)
|
|
|
|
eda_intraday_features.reset_index(inplace=True)
|
|
|
|
return eda_intraday_features
|
|
|
|
|
|
def dbdp_features(sensor_data_files, time_segment, provider, filter_data_by_segment, *args, **kwargs):
|
|
eda_intraday_data = pd.read_csv(sensor_data_files["sensor_data"])
|
|
|
|
requested_intraday_features = provider["FEATURES"]
|
|
# name of the features this function can compute
|
|
base_intraday_features_names = ["maxeda", "mineda", "avgeda", "medianeda", "modeeda", "stdeda", "diffmaxmodeeda",
|
|
"diffminmodeeda", "entropyeda"]
|
|
# the subset of requested features this function can compute
|
|
intraday_features_to_compute = list(set(requested_intraday_features) & set(base_intraday_features_names))
|
|
|
|
# extract features from intraday data
|
|
eda_intraday_features = extractEDAFeaturesFromIntradayData(eda_intraday_data,
|
|
intraday_features_to_compute, time_segment,
|
|
filter_data_by_segment)
|
|
|
|
return eda_intraday_features |