Module minder_utils.scripts.feature_selection
Expand source code
from minder_utils.models.feature_selectors.supervised.filter import Supervised_Filter
from minder_utils.models.feature_selectors.unsupervised.filter import Unsupervised_Filter
from prettytable import PrettyTable
from minder_utils.evaluate.eval_utils import split_by_ids, get_scores
from minder_utils.formatting.format_util import format_mean_std
import pandas as pd
class Feature_Selection:
'''
This class can help you to choose the best feature selection method for your model.
To run the script, your model must implement ```reset()``` method, which is to
reset the parameters of your model for testing.
```Example```
```
from minder_utils.scripts.feature_selection import Feature_Selection
from minder_utils.models.classifiers import Classifiers
classifier = Classifiers('nn')
feature_select = Feature_Selection(classifier)
feature_select.evaluate(X, y, p_ids, num_runs=10)
```
'''
def __init__(self, model, proportion=90):
self.model = model
self.feature_selector = [Supervised_Filter(proportion=proportion), Unsupervised_Filter()]
def evaluate(self, X, y, p_ids, num_runs=10):
header = ['Supervision', 'Method', 'Feature selected ({} in total)'.format(X.shape[1]), 'sensitivity', 'specificity', 'acc', 'f1']
results = []
for selector in self.feature_selector:
sen, spe, accs, f1s, num_feats = [], [], [], [], []
info = selector.get_info()
for model_type in info:
for run in range(num_runs):
X_train, y_train, X_test, y_test = split_by_ids(X, y, p_ids, seed=run)
self.model.reset()
selector.reset_model(model_type)
selector.fit(X_train, y_train)
X_train_trans = selector.transform(X_train)
X_test_trans = selector.transform(X_test)
self.model.fit(X_train_trans, y_train)
sensitivity, specificity, acc, f1 = get_scores(y_test, self.model.predict(X_test_trans))
if sensitivity is not None and str(sensitivity) != 'nan':
sen.append(sensitivity)
spe.append(specificity)
accs.append(acc)
f1s.append(f1)
num_feats.append(X_train_trans.shape[1])
row = [selector.__name__()[0], selector.__name__()[1], format_mean_std(num_feats),
format_mean_std(sen), format_mean_std(spe), format_mean_std(accs), format_mean_std(f1s)]
results.append(row)
df_results = pd.DataFrame(results, columns=header)
return df_resultsClasses
class Feature_Selection (model, proportion=90)-
This class can help you to choose the best feature selection method for your model.
To run the script, your model must implement
reset()method, which is to reset the parameters of your model for testing.Examplefrom minder_utils.scripts.feature_selection import Feature_Selection from minder_utils.models.classifiers import Classifiers classifier = Classifiers('nn') feature_select = Feature_Selection(classifier) feature_select.evaluate(X, y, p_ids, num_runs=10)Expand source code
class Feature_Selection: ''' This class can help you to choose the best feature selection method for your model. To run the script, your model must implement ```reset()``` method, which is to reset the parameters of your model for testing. ```Example``` ``` from minder_utils.scripts.feature_selection import Feature_Selection from minder_utils.models.classifiers import Classifiers classifier = Classifiers('nn') feature_select = Feature_Selection(classifier) feature_select.evaluate(X, y, p_ids, num_runs=10) ``` ''' def __init__(self, model, proportion=90): self.model = model self.feature_selector = [Supervised_Filter(proportion=proportion), Unsupervised_Filter()] def evaluate(self, X, y, p_ids, num_runs=10): header = ['Supervision', 'Method', 'Feature selected ({} in total)'.format(X.shape[1]), 'sensitivity', 'specificity', 'acc', 'f1'] results = [] for selector in self.feature_selector: sen, spe, accs, f1s, num_feats = [], [], [], [], [] info = selector.get_info() for model_type in info: for run in range(num_runs): X_train, y_train, X_test, y_test = split_by_ids(X, y, p_ids, seed=run) self.model.reset() selector.reset_model(model_type) selector.fit(X_train, y_train) X_train_trans = selector.transform(X_train) X_test_trans = selector.transform(X_test) self.model.fit(X_train_trans, y_train) sensitivity, specificity, acc, f1 = get_scores(y_test, self.model.predict(X_test_trans)) if sensitivity is not None and str(sensitivity) != 'nan': sen.append(sensitivity) spe.append(specificity) accs.append(acc) f1s.append(f1) num_feats.append(X_train_trans.shape[1]) row = [selector.__name__()[0], selector.__name__()[1], format_mean_std(num_feats), format_mean_std(sen), format_mean_std(spe), format_mean_std(accs), format_mean_std(f1s)] results.append(row) df_results = pd.DataFrame(results, columns=header) return df_resultsMethods
def evaluate(self, X, y, p_ids, num_runs=10)-
Expand source code
def evaluate(self, X, y, p_ids, num_runs=10): header = ['Supervision', 'Method', 'Feature selected ({} in total)'.format(X.shape[1]), 'sensitivity', 'specificity', 'acc', 'f1'] results = [] for selector in self.feature_selector: sen, spe, accs, f1s, num_feats = [], [], [], [], [] info = selector.get_info() for model_type in info: for run in range(num_runs): X_train, y_train, X_test, y_test = split_by_ids(X, y, p_ids, seed=run) self.model.reset() selector.reset_model(model_type) selector.fit(X_train, y_train) X_train_trans = selector.transform(X_train) X_test_trans = selector.transform(X_test) self.model.fit(X_train_trans, y_train) sensitivity, specificity, acc, f1 = get_scores(y_test, self.model.predict(X_test_trans)) if sensitivity is not None and str(sensitivity) != 'nan': sen.append(sensitivity) spe.append(specificity) accs.append(acc) f1s.append(f1) num_feats.append(X_train_trans.shape[1]) row = [selector.__name__()[0], selector.__name__()[1], format_mean_std(num_feats), format_mean_std(sen), format_mean_std(spe), format_mean_std(accs), format_mean_std(f1s)] results.append(row) df_results = pd.DataFrame(results, columns=header) return df_results