def run_lda():
kfold = KFold(n_splits=5, shuffle=True, random_state=1)
processor = P1()
datasets = [Diabetes(), Adult()]
estimators = [
Config(name='lda',
estimator=LinearDiscriminantAnalysis(),
cv=kfold,
params={})
]
for dataset in datasets:
for estimator in estimators:
estimator = processor.get_default_model(dataset=dataset,
estimator=estimator)
processor.process_validations(dataset=dataset, estimator=estimator)
processor.plot_validation()
for dataset in datasets:
for estimator in estimators:
estimator = processor.get_default_model(dataset=dataset,
estimator=estimator)
processor.param_selection(dataset=dataset, estimator=estimator)
processor.print_best_params()
for dataset in datasets:
for estimator in estimators:
processor.process(dataset=dataset, estimator=estimator)
processor.plot_learning_curves()
def run_nn_opt():
kfold = KFold(n_splits=5, shuffle=True, random_state=1)
processor = P1()
datasets = [Diabetes()]
# 'random_hill_climb', 'simulated_annealing', 'genetic_alg'
estimators = [
Config(name='NN_%s' % title('random_hill_climb'),
estimator=mlrose.NeuralNetwork(
algorithm='random_hill_climb',
random_state=1,
max_iters=200,
hidden_nodes=[64],
early_stopping=True,
),
cv=kfold,
params={'restarts': [0, 10, 20, 30, 40, 50]}),
Config(name='NN_%s' % title('simulated_annealing'),
estimator=mlrose.NeuralNetwork(
algorithm='simulated_annealing',
random_state=1,
max_iters=200,
hidden_nodes=[64],
early_stopping=True,
),
cv=kfold,
params={'max_iters': [200]}),
Config(name='NN_%s' % title('genetic_alg'),
estimator=mlrose.NeuralNetwork(
algorithm='genetic_alg',
random_state=1,
max_iters=200,
hidden_nodes=[64],
early_stopping=True,
),
cv=kfold,
params={
'pop_size': [100, 200, 300, 400, 500, 600, 700, 800, 900],
'mutation_prob':
[0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9]
}),
]
for dataset in datasets:
for estimator in estimators:
estimator = processor.get_default_model(dataset=dataset,
estimator=estimator)
processor.process_validations(dataset=dataset, estimator=estimator)
processor.plot_validation()
for dataset in datasets:
for estimator in estimators:
estimator = processor.get_default_model(dataset=dataset,
estimator=estimator)
processor.param_selection(dataset=dataset, estimator=estimator)
processor.print_best_params()
for dataset in datasets:
for estimator in estimators:
processor.process(dataset=dataset, estimator=estimator)
processor.plot_learning_curves()
def run_nn_opt():
kfold = KFold(n_splits=5, shuffle=True, random_state=1)
processor = P1()
datasets = [Diabetes()]
estimators = [
Config(name='NN_%s' % title('pca'),
estimator=Pipeline([('pca', PCA(n_components=7,
random_state=0)),
('nn', MLPClassifier(max_iter=1000))]),
cv=kfold,
params={}),
Config(name='NN_%s' % title('ica'),
estimator=Pipeline([('pca',
FastICA(n_components=7, random_state=0)),
('nn', MLPClassifier(max_iter=1000))]),
cv=kfold,
params={}),
Config(name='NN_%s' % title('rp'),
estimator=Pipeline([('pca',
SparseRandomProjection(random_state=0,
n_components=7)),
('nn', MLPClassifier(max_iter=1000))]),
cv=kfold,
params={}),
Config(name='NN_%s' % title('dl'),
estimator=Pipeline([
('pca',
MiniBatchDictionaryLearning(n_components=7,
batch_size=200,
n_iter=1000,
random_state=0)),
('nn', MLPClassifier(max_iter=1000))
]),
cv=kfold,
params={})
]
for dataset in datasets:
for estimator in estimators:
processor.process(dataset=dataset,
estimator=estimator,
control=True)
processor.plot_learning_curves(control=True)
def run_feature_importance():
processor = Processor3()
processor.latext_start_figure()
for dataset in [Diabetes(), Adult()]:
processor = Processor3()
processor.latext_start_figure()
X_train, X_test, y_train, y_test, _ = dataset.get_data(model='KMeans')
forest = RandomForestClassifier(n_estimators=500, random_state=1)
forest.fit(X_train, y_train)
importances = forest.feature_importances_
indices = np.argsort(importances)[::-1]
top_10 = []
top_10_vals = []
top_10_idx = []
for f, g in zip(range(X_train.shape[1]), indices[:10]):
print("%2d) % -*s %f" % (f + 1, 30, dataset.fields[indices[f]],
importances[indices[f]]))
top_10.append(dataset.fields[indices[f]])
top_10_idx.append(indices[f])
top_10_vals.append(importances[indices[f]])
print(top_10)
print(top_10_idx)
plt.title('Feature Importance')
plt.bar(top_10, top_10_vals, align='center')
plt.xticks(top_10, rotation=90)
plt.tight_layout()
# plt.show()
filename = '%s_%s' % ('features', dataset.__class__.__name__)
chart_path = 'report/images/%s.png' % filename
plt.savefig(chart_path)
plt.close()
processor.latex_subgraph(dataset=dataset.__class__.__name__,
fig=filename,
caption=dataset.__class__.__name__,
filename=filename)
processor.latex_end_figure(caption=f"Feature Importance",
fig=f"feature_importance")
def run_nn_opt_clusters():
kfold = KFold(n_splits=5, shuffle=True, random_state=1)
processor = P1()
datasets = [Diabetes()]
dr_models = [
PCA(n_components=7, random_state=0),
FastICA(n_components=7, random_state=0),
MiniBatchDictionaryLearning(n_components=7,
batch_size=200,
n_iter=1000,
random_state=0),
SparseRandomProjection(random_state=0, n_components=7)
]
clustering_models = []
for i in [2]:
clustering_models.append(
KMeans(n_clusters=i,
init='k-means++',
n_init=10,
max_iter=600,
random_state=0,
tol=0.0001))
clustering_models.append(
GaussianMixture(n_components=i,
n_init=10,
max_iter=600,
random_state=0,
tol=0.0001))
configs = {}
for model in dr_models:
configs[model.__class__.__name__.lower()] = {}
for dataset in datasets:
dataset_name = dataset.__class__.__name__
X_train, X_test, y_train, y_test, target_names = dataset.get_data(
model='KMeans')
for model in dr_models:
reduction = model.__class__.__name__.lower()
print(reduction)
if reduction == 'lineardiscriminantanalysis':
X_r = model.fit_transform(X_train, y_train)
X_r_test = model.transform(X_test)
else:
X_r = model.fit_transform(X_train)
X_r_test = model.transform(X_test)
for clustering_model in clustering_models:
try:
clusters = clustering_model.n_components
except:
clusters = clustering_model.n_clusters
preds = clustering_model.fit_predict(X_train)
preds_test = clustering_model.predict(X_test)
df = pd.DataFrame(X_r)
df['cluster'] = preds
df_test = pd.DataFrame(X_r_test)
df_test['cluster'] = preds_test
print('done')
data = {
'X_train': df,
'X_test': df_test,
'y_train': y_train,
'y_test': y_test
}
estimator = Config(
name='NN_%s_%s_%i' %
(clustering_model.__class__.__name__.lower(), reduction,
clusters),
estimator=Pipeline([('nn', MLPClassifier(max_iter=1000))]),
cv=kfold,
params={'nn__restarts': [10]})
processor.process(dataset=dataset,
estimator=estimator,
control=True,
data=data)
processor.plot_learning_curves(control=True, suffix='_part5')
except OSError as error:
print("Directory '%s' can not be created")
parser = argparse.ArgumentParser(description='Find X Coding Quiz')
parser.add_argument('-m',
'--mode',
help='Mode',
default='debug',
dest='mode')
args = parser.parse_args()
processor = Processor3()
datasets = [
Diabetes(),
Adult(),
]
run_feature_importance()
compute_kmeans_elbow_curves()
visualize_kmeans_clusters()
run_dimension_reductions()
run_nn_opt()
run_nn_opt_clusters()
## TODO: Uncomment to get the other charts
compute_kmeans_elbow_curves()
compute_em_elbow_curves()
run_bics()
from sklearn.decomposition import MiniBatchDictionaryLearning
from sklearn.decomposition import PCA
from sklearn.metrics import adjusted_rand_score
from sklearn.metrics import silhouette_score
from sklearn.mixture import GaussianMixture
from sklearn.pipeline import Pipeline
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
from sklearn.random_projection import SparseRandomProjection
from hw1.utils import Adult
from hw1.utils import Diabetes
from hw3.main import Processor3
datasets = [Adult(), Diabetes()]
for dataset in datasets:
processor = Processor3()
processor.latext_start_figure()
X_train, X_test, y_train, y_test, _ = dataset.get_data(model='KMeans')
n_clusters = len(dataset.label_encoder.classes_)
pca = PCA(n_components=0.95)
pca.fit(X_train)
n_components = pca.components_.shape[0]
print(f"n_components: {n_components}")
dr_models = [
PCA(n_components=n_components, random_state=0),
FastICA(n_components=n_components, random_state=0),
MiniBatchDictionaryLearning(n_components=n_components,
alpha=1,
def run_dimension_reductions():
global mean
for dataset in [Diabetes(), Adult()]:
processor = Processor3()
processor.latext_start_figure()
X_train, X_test, y_train, y_test, _ = dataset.get_data(model='KMeans')
pca = PCA(n_components=0.95)
pca.fit(X_train)
n_components = pca.components_.shape[0]
print(f"n_components: {n_components}")
whiten = True
random_state = 0
dr_models = [
PCA(n_components=n_components, random_state=0),
FastICA(n_components=n_components, random_state=0),
MiniBatchDictionaryLearning(n_components=n_components,
alpha=1,
batch_size=200,
n_iter=10,
random_state=random_state),
SparseRandomProjection(random_state=0, n_components=n_components)
]
for pca in dr_models:
X_train = pd.DataFrame(X_train)
y_train = pd.DataFrame(y_train)
if isinstance(pca, SparseRandomProjection):
X_train_PCA = pca.fit_transform(X_train)
X_train_PCA = pd.DataFrame(data=X_train_PCA,
index=X_train.index)
X_train_PCA_inverse = np.array(X_train_PCA).dot(
pca.components_.todense())
X_train_PCA_inverse = pd.DataFrame(data=X_train_PCA_inverse,
index=X_train.index)
scatterPlot(X_train_PCA, y_train, pca.__class__.__name__)
elif isinstance(pca, MiniBatchDictionaryLearning):
X_train_PCA = pca.fit_transform(X_train)
X_train_PCA = pd.DataFrame(data=X_train_PCA,
index=X_train.index)
X_train_PCA_inverse = np.array(X_train_PCA).dot(
pca.components_) + np.array(X_train.mean(axis=0))
X_train_PCA_inverse = pd.DataFrame(data=X_train_PCA_inverse,
index=X_train.index)
scatterPlot(X_train_PCA, y_train, pca.__class__.__name__)
else:
X_train_PCA = pca.fit_transform(X_train)
X_train_PCA = pd.DataFrame(data=X_train_PCA,
index=X_train.index)
X_train_PCA_inverse = pca.inverse_transform(X_train_PCA)
X_train_PCA_inverse = pd.DataFrame(data=X_train_PCA_inverse,
index=X_train.index)
scatterPlot(X_train_PCA, y_train, pca.__class__.__name__)
# plt.show()
anomalyScoresPCA = anomalyScores(X_train, X_train_PCA_inverse)
mean = np.mean(anomalyScoresPCA)
print(mean)
preds = plotResults(y_train, anomalyScoresPCA, True,
pca.__class__.__name__,
dataset.__class__.__name__, mean)
processor.latex_end_figure(
caption=f"{dataset.__class__.__name__} Precision-Recall Curve",
fig=f"pr_{dataset.__class__.__name__}")