diff --git a/config.yaml b/config.yaml index f03ca101..202abe48 100644 --- a/config.yaml +++ b/config.yaml @@ -688,7 +688,7 @@ ALL_CLEANING_INDIVIDUAL: COMPUTE: True IMPUTE_PHONE_SELECTED_EVENT_FEATURES: COMPUTE: False - TYPE: median # options: zero, mean, median or k-nearest + TYPE: zero # options: zero, mean, median or k-nearest MIN_DATA_YIELDED_MINUTES_TO_IMPUTE: 0.33 COLS_NAN_THRESHOLD: 1 # set to 1 remove only columns that contains all NaN COLS_VAR_THRESHOLD: True @@ -723,7 +723,7 @@ ALL_CLEANING_OVERALL: COMPUTE: True IMPUTE_PHONE_SELECTED_EVENT_FEATURES: COMPUTE: False - TYPE: median # options: zero, mean, median or k-nearest + TYPE: zero # options: zero, mean, median or k-nearest MIN_DATA_YIELDED_MINUTES_TO_IMPUTE: 0.33 COLS_NAN_THRESHOLD: 1 # set to 1 remove only columns that contains all NaN COLS_VAR_THRESHOLD: True diff --git a/src/features/all_cleaning_individual/straw/main.py b/src/features/all_cleaning_individual/straw/main.py index f2c1bcd0..19ae71f5 100644 --- a/src/features/all_cleaning_individual/straw/main.py +++ b/src/features/all_cleaning_individual/straw/main.py @@ -4,25 +4,26 @@ import math, sys import yaml from sklearn.impute import KNNImputer +from sklearn.preprocessing import StandardScaler import matplotlib.pyplot as plt import seaborn as sns def straw_cleaning(sensor_data_files, provider): features = pd.read_csv(sensor_data_files["sensor_data"][0]) + esm_cols = features.loc[:, features.columns.str.startswith('phone_esm_straw')] # Get target (esm) columns with open('config.yaml', 'r') as stream: config = yaml.load(stream, Loader=yaml.FullLoader) - #Filter-out all rows that do not have the target column available + # (1) FILTER_OUT THE ROWS THAT DO NOT HAVE THE TARGET COLUMN AVAILABLE if config['PARAMS_FOR_ANALYSIS']['TARGET']['COMPUTE']: target = config['PARAMS_FOR_ANALYSIS']['TARGET']['LABEL'] # get target label from config features = features[features['phone_esm_straw_' + target].notna()].reset_index() test_cols = [col for col in features.columns if 'phone_calls' in col or 'phone_messages' in col] - # TODO: reorder the cleaning steps so it makes sense for the analysis # TODO: add conditions that differentiates cleaning steps for standardized and nonstandardized features, for this # the snakemake rules will also have to come with additional parameter (in rules/features.smk) @@ -36,7 +37,7 @@ def straw_cleaning(sensor_data_files, provider): # because of the lack of the availability. Secondly, there's a high importance that features data frame is checked if and NaN # values still exist. - # Impute selected features event + # (2) PARTIAL IMPUTATION: IMPUTE DATA DEPENDEND ON THE FEATURES GROUP (e.g., phone or E4 features) impute_phone_features = provider["IMPUTE_PHONE_SELECTED_EVENT_FEATURES"] if impute_phone_features["COMPUTE"]: if not 'phone_data_yield_rapids_ratiovalidyieldedminutes' in features.columns: @@ -55,7 +56,7 @@ def straw_cleaning(sensor_data_files, provider): mask = features['phone_data_yield_rapids_ratiovalidyieldedminutes'] > impute_phone_features['MIN_DATA_YIELDED_MINUTES_TO_IMPUTE'] features.loc[mask, phone_cols] = impute(features[mask][phone_cols], method=impute_phone_features["TYPE"].lower()) - # Drop rows with the value of data_yield_column less than data_yield_ratio_threshold + # ??? Drop rows with the value of data_yield_column less than data_yield_ratio_threshold ??? data_yield_unit = provider["DATA_YIELD_FEATURE"].split("_")[3].lower() data_yield_column = "phone_data_yield_rapids_ratiovalidyielded" + data_yield_unit @@ -65,10 +66,10 @@ def straw_cleaning(sensor_data_files, provider): if provider["DATA_YIELD_RATIO_THRESHOLD"]: features = features[features[data_yield_column] >= provider["DATA_YIELD_RATIO_THRESHOLD"]] - # Remove cols if threshold of NaN values is passed (should be <= if even all NaN columns must be preserved) + # (3) REMOVE COLS IF THEIR NAN THRESHOLD IS PASSED (should be <= if even all NaN columns must be preserved) features = features.loc[:, features.isna().sum() < provider["COLS_NAN_THRESHOLD"] * features.shape[0]] - # Remove cols where variance is 0 + # (4) REMOVE COLS WHERE VARIANCE IS 0 if provider["COLS_VAR_THRESHOLD"]: features.drop(features.std()[features.std() == 0].index.values, axis=1, inplace=True) @@ -77,7 +78,7 @@ def straw_cleaning(sensor_data_files, provider): if esm not in features: features[esm] = esm_cols[esm] - # Drop highly correlated features - To-Do še en thershold var, ki je v config + kako se tretirajo NaNs? + # (5) DROP HIGHLY CORRELATED FEATURES drop_corr_features = provider["DROP_HIGHLY_CORRELATED_FEATURES"] if drop_corr_features["COMPUTE"]: @@ -98,7 +99,26 @@ def straw_cleaning(sensor_data_files, provider): sns.set(rc={"figure.figsize":(16, 8)}) sns.heatmap(features.isna(), cbar=False) - plt.savefig(f'features_nans.png', bbox_inches='tight') + plt.savefig(f'features_nans_bf_knn.png', bbox_inches='tight') + + # KNN IMPUTATION + features = impute(features, method="knn") + + sns.set(rc={"figure.figsize":(16, 8)}) + sns.heatmap(features.isna(), cbar=False) + plt.savefig(f'features_nans_af_knn.png', bbox_inches='tight') + + + ## STANDARDIZATION - should it happen before or after kNN imputation? + # TODO: check if there are additional columns that need to be excluded from the standardization + excluded_columns = ['local_segment', 'local_segment_label', 'local_segment_start_datetime', 'local_segment_end_datetime'] + excluded_columns += [col for col in features.columns if "level_1" in col] + + features.loc[:, ~features.columns.isin(excluded_columns)] = StandardScaler().fit_transform(features.loc[:, ~features.columns.isin(excluded_columns)]) + + # VERIFY IF THERE ARE ANY NANS LEFT IN THE DATAFRAME + if features.isna.any().any(): + raise ValueError sys.exit() @@ -106,14 +126,14 @@ def straw_cleaning(sensor_data_files, provider): def impute(df, method='zero'): - def k_nearest(df): # TODO: if needed, implement k-nearest imputation / interpolation + def k_nearest(df): imputer = KNNImputer(n_neighbors=3) return pd.DataFrame(imputer.fit_transform(df), columns=df.columns) - return { # rest of the columns should be imputed with the selected method + return { 'zero': df.fillna(0), 'mean': df.fillna(df.mean()), 'median': df.fillna(df.median()), - 'k-nearest': k_nearest(df) + 'knn': k_nearest(df) }[method]