def run():
candidates = CandidateFeatureVector.objects.all().values()
candidates = DataUtil.remove_fake_candidates(candidates)
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
candidates_df.drop(columns=['candidate_id'], inplace=True)
candidates_df = EncodingUtil.basic_label_encode_cols(
candidates_df, ConstantsUtil.BASIC_ENCODE_COLS)
candidates_df = EncodingUtil.sort_position_cols_and_encode(
candidates_df, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
candidates_df.drop(
columns=ConstantsUtil.FEATURES_IGNORED_BY_INFORMATION_GAIN,
inplace=True)
X = candidates_df.drop(columns=['classification'])
y = candidates_df['classification']
knn = KerasNN(input_dim=len(X.columns))
X_train, X_test, y_train, y_test = knn.fit_model(X, y, 120, .3, .2)
knn.plot_training_validation()
knn.plot_roc_curve()
print(knn.get_classification_report(X_test, y_test))
knn.get_accuracy_stats(X_test, y_test)
results = knn.test_n_iterations(X, y, 120, .3, .2, 10)
print(results)
'''
def run():
candidates = CandidateFeatureVector.objects.all().values()
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
candidates_df.drop(columns=['candidate_id', 'classification'], inplace=True)
candidates_df = EncodingUtil.basic_label_encode_cols(candidates_df, ConstantsUtil.BASIC_ENCODE_COLS)
candidates_df = EncodingUtil.sort_position_cols_and_encode(candidates_df, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
pca = PCA(.9)
pca.fit(candidates_df)
print(pca.components_)
def get_candidates_df(candidates, drop_fakes=False, convert_fakes=True, drop_min_features=False):
if drop_fakes:
candidates = DataUtil.remove_fake_candidates(candidates)
elif convert_fakes:
candidates = DataUtil.convert_fake_candidates(candidates)
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
candidates_df.drop(columns=['candidate_id'], inplace=True)
candidates_df = EncodingUtil.basic_label_encode_cols(candidates_df, ConstantsUtil.BASIC_ENCODE_COLS)
candidates_df = EncodingUtil.sort_position_cols_and_encode(candidates_df, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
if drop_min_features:
candidates_df.drop(columns=ConstantsUtil.FEATURES_IGNORED_BY_INFORMATION_GAIN, inplace=True)
return candidates_df
def run():
candidates = CandidateFeatureVector.objects.all().values()
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
candidates_df.drop(columns=['candidate_id'], inplace=True)
candidates_df = EncodingUtil.basic_label_encode_cols(
candidates_df, ConstantsUtil.BASIC_ENCODE_COLS)
candidates_df = EncodingUtil.sort_position_cols_and_encode(
candidates_df, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
dt = DecisionTree(criterion="entropy")
dt.set_data(candidates_df, 'classification')
X_train, X_test, y_train, y_test = dt.split_test_data(.3, True)
dt.get_roc_curve(X_train, X_test, y_train, y_test)
dt.fit_and_predict(X_train, X_test, y_train)
print(dt.get_confusion_matrix(y_test))
print(dt.get_classification_report(y_test))
dt.get_learning_curve()
def run():
candidates = CandidateFeatureVector.objects.all().values()
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
candidates_df.drop(columns=['candidate_id'], inplace=True)
candidates_df = EncodingUtil.basic_label_encode_cols(
candidates_df, ConstantsUtil.BASIC_ENCODE_COLS)
candidates_df = EncodingUtil.sort_position_cols_and_encode(
candidates_df, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
svm = SVM(C=.75, kernel='poly')
X_train, X_test, y_train, y_test = svm.split_test_data(
candidates_df, .3, 'classification', True)
svm.fit_and_predict(X_train, X_test, y_train)
print(svm.get_confusion_matrix(y_test))
print(svm.get_classification_report(y_test))
estimator = svm.get_model()
selector = GeneticSelectionCV(estimator,
cv=5,
verbose=1,
scoring="accuracy",
max_features=50,
n_population=50,
crossover_proba=0.5,
mutation_proba=0.2,
n_generations=40,
crossover_independent_proba=0.5,
mutation_independent_proba=0.05,
tournament_size=3,
n_gen_no_change=10,
caching=True,
n_jobs=-1)
X, y = svm.get_data()
selector = selector.fit(X, y)
print(selector.support_)
def run():
feat_importances = []
candidates = CandidateFeatureVector.objects.all().values()
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
for i in range(0,10):
y = candidates_df['classification']
X = candidates_df.copy()
X.drop(columns=['candidate_id', 'classification'], inplace=True)
X = EncodingUtil.basic_label_encode_cols(X, ConstantsUtil.BASIC_ENCODE_COLS)
X = EncodingUtil.sort_position_cols_and_encode(X, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
importances = mutual_info_classif(X.values, y)
feat_importances.append(pd.Series(importances, X.columns))
feature_df = pd.DataFrame(feat_importances)
importance_means = feature_df.mean()
importance_means.plot(kind='barh')
plt.show()
print(importance_means)
def run():
candidates = CandidateFeatureVector.objects.all().values()
candidates_df = pd.DataFrame(candidates)
candidates_df.set_index('id', inplace=True)
candidates_df.drop(columns=['candidate_id'], inplace=True)
candidates_df = EncodingUtil.basic_label_encode_cols(
candidates_df, ConstantsUtil.BASIC_ENCODE_COLS)
candidates_df = EncodingUtil.sort_position_cols_and_encode(
candidates_df, ConstantsUtil.STRING_TUPLE_ENCODE_COLS)
gnb = GaussianNB()
gnb.set_data(candidates_df, 'classification')
X_train, X_test, y_train, y_test = gnb.split_test_data(.3, True)
gnb.get_roc_curve(X_train, X_test, y_train, y_test)
gnb.fit_and_predict(X_train, X_test, y_train)
print(gnb.get_confusion_matrix(y_test))
print(gnb.get_classification_report(y_test))
gnb.get_learning_curve()
print(gnb.get_avg_metrics_for_n_iterations(10, .3, True))