def fit_transform(self, X, y): predictions, scores = cv_predict_transform(self.model_estimators, X, y, self.prediction_function, self.verbose) self.scores = scores self.predictions = predictions return predictions if __name__ == "__main__": from prepare_data import subjects, get_nii_data, load_stimuli _, _, stimuli = load_stimuli() data = get_nii_data(subjects[0]) from multi_select_k_best import MultiSelectKBest from sklearn.feature_selection import SelectKBest from sklearn.feature_selection import f_classif from sklearn.linear_model import LogisticRegression from sklearn.pipeline import Pipeline ## A first pipeline using globally selected features ## and l2 logistic regression selector = MultiSelectKBest(f_classif, k=500) estimators = [LogisticRegression(C=C, penalty='l2') for C in 2. ** np.arange(-24, 0, 2)]
import numpy as np if __name__ == "__main__": # re-generate layer 1 results from layer_1_predictor import CvPredictTransform from prepare_data import subjects, get_nii_data, load_stimuli _, _, stimuli = load_stimuli() data = get_nii_data(subjects[2]) from multi_select_k_best import MultiSelectKBest from sklearn.feature_selection import SelectKBest from sklearn.feature_selection import f_classif from sklearn.linear_model import LogisticRegression from sklearn.svm import SVC from sklearn.pipeline import Pipeline ## A first pipeline using globally selected features ## and l2 logistic regression selector = MultiSelectKBest(f_classif, k=100) estimators = [LogisticRegression(C=C, penalty='l2') for C in [10]] #]2. ** np.arange(-24, 0, 2)] # estimators = [SVC(C=C, probability=True, kernel="linear") # for C in 2. ** np.arange(-24, 2, 2)] first_layer_predictor = CvPredictTransform(model_estimators=estimators) pipeline = Pipeline([('feature_reduction', selector), ('first_layer_prediction', first_layer_predictor)])
def fit_transform(self, X, y): predictions, scores = cv_predict_transform(self.model_estimators, X, y, self.prediction_function, self.verbose) self.scores = scores self.predictions = predictions return predictions if __name__ == "__main__": from prepare_data import subjects, get_nii_data, load_stimuli _, _, stimuli = load_stimuli() data = get_nii_data(subjects[0]) from multi_select_k_best import MultiSelectKBest from sklearn.feature_selection import SelectKBest from sklearn.feature_selection import f_classif from sklearn.linear_model import LogisticRegression from sklearn.pipeline import Pipeline scaler = StandardScaler() ## A first pipeline using globally selected features ## and l2 logistic regression selector = MultiSelectKBest(f_classif, k=300) estimators = [ LogisticRegression(C=C, penalty='l2', intercept_scaling=100)
import numpy as np if __name__ == "__main__": # re-generate layer 1 results from layer_1_predictor import CvPredictTransform from prepare_data import subjects, get_nii_data, load_stimuli _, _, stimuli = load_stimuli() data = get_nii_data(subjects[2]) from multi_select_k_best import MultiSelectKBest from sklearn.feature_selection import SelectKBest from sklearn.feature_selection import f_classif from sklearn.linear_model import LogisticRegression from sklearn.svm import SVC from sklearn.pipeline import Pipeline ## A first pipeline using globally selected features ## and l2 logistic regression selector = MultiSelectKBest(f_classif, k=100) estimators = [LogisticRegression(C=C, penalty='l2') for C in [10]]#]2. ** np.arange(-24, 0, 2)] # estimators = [SVC(C=C, probability=True, kernel="linear") # for C in 2. ** np.arange(-24, 2, 2)] first_layer_predictor = CvPredictTransform(model_estimators=estimators) pipeline = Pipeline([('feature_reduction', selector),