%matplotlib inline
import matplotlib as plt
import numpy as np
from autotagger.stackoverflow.preprocess import load_pickle_sklearn_format
from sklearn import cross_validation,linear_model
from sklearn.datasets import make_multilabel_classification
from sklearn.metrics import f1_score, precision_score, recall_score
from sklearn.multiclass import OneVsRestClassifier
import pickle
X,Y = load_pickle_sklearn_format("1GB_100_features")
X_train, X_test, Y_train, Y_test = cross_validation.train_test_split(X,Y,test_size=0.80, random_state=42)
X_train.shape, X_test.shape, Y_train.shape, Y_test.shape
((166811, 100), (667245, 100), (166811, 100), (667245, 100))
clf = linear_model.LinearRegression()
meta_clf = OneVsRestClassifier(clf)
meta_clf.fit(X_train,Y_train)
OneVsRestClassifier(estimator=LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False), n_jobs=1)
Y_pred = meta_clf.predict(X_test)
# macro average refers to the average f1_score for each label
f1_score(Y_test,Y_pred, average='macro')
/home/felipe/auto-tagger/venv3/lib/python3.4/site-packages/sklearn/metrics/classification.py:1074: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples. 'precision', 'predicted', average, warn_for)
0.018744593095539479
# if we just consider the labels that have had at least one instance predicted,
# our score goes up:
label_scores = f1_score(Y_test,Y_pred,average=None)
valid_label_indices = np.nonzero(label_scores)[0]
f1_score(Y_test,Y_pred,average='macro',labels=valid_label_indices)
/home/felipe/auto-tagger/venv3/lib/python3.4/site-packages/sklearn/metrics/classification.py:1074: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples. 'precision', 'predicted', average, warn_for)
0.23430741369424352
# micro average refers to the average f1_score for each instance
f1_score(Y_test,Y_pred,average='micro')
0.13040891760528536