from sklearn.naive_bayes import GaussianNB
import personal_settings
import SDK as sdk
import os
path = personal_settings.PATH
algorithm = os.path.basename(__file__).split(".py")[0]
# execute for all datasets:
for extraction_type in os.listdir(path):
print("Starting new classification. Extraction method : " + extraction_type)
folder = path + extraction_type + "/"
try:
clf = GaussianNB()
sdk.evaluate_classifier(clf, folder, extraction_type, algorithm)
except Exception as e:
print(str(e))
pass
print("Ended extraction : " + extraction_type)
#execute for all datasets:
for extraction_type in datasets:
print("Starting new classification. Extraction method : " +
extraction_type)
folder = path + extraction_type + "/"
for weights in ['distance']:
for n_neighbors in [3, 5, 6, 7]:
print("Starting new classification. Extraction method : " +
extraction_type)
print("weights = " + weights)
try:
clf = neighbors.KNeighborsClassifier(n_neighbors=n_neighbors,
n_jobs=-1,
weights=weights)
sdk.evaluate_classifier(clf,
folder,
extraction_type,
algorithm,
suffixe="_" + str(n_neighbors) +
"NN_" + weights)
except Exception as e:
print(str(e))
pass
print("Ended extraction : " + extraction_type)
algorithm = os.path.basename(__file__).split(".py")[0]
datasets = os.listdir(path)
datasets = personal_settings.LARGE_DATASETS
# execute for all datasets:
for extraction_type in datasets:
print("Starting new classification. Extraction method : " +
extraction_type)
folder = path + extraction_type + "/"
n_trees = 100
try:
in_clf_mlp = MLPClassifier(verbose=True,
hidden_layer_sizes=300,
early_stopping=True)
in_clf_rd = ensemble.RandomForestClassifier(n_jobs=-1,
verbose=20,
n_estimators=n_trees)
clf = OneVsRestClassifier(in_clf_mlp, n_jobs=-1)
sdk.evaluate_classifier(clf,
folder,
extraction_type,
algorithm,
suffixe="_mlp")
except Exception as e:
print(str(e))
pass
print("Ended extraction : " + extraction_type)
import personal_settings
from sklearn import ensemble
import os
import SDK as sdk
path = personal_settings.PATH
algorithm = os.path.basename(__file__).split(".py")[0]
datasets = ["MSD-MARSYAS"]
# datasets = os.listdir(path)
for extraction_type in datasets:
print("Starting new classification. Extraction method : " +
extraction_type)
folder = path + extraction_type + "/"
try:
for n_trees in [100]:
clf = ensemble.RandomForestClassifier(n_jobs=-1,
verbose=20,
n_estimators=n_trees)
sdk.evaluate_classifier(clf,
folder,
extraction_type,
algorithm,
suffixe="_" + str(n_trees) + "_trees")
except Exception as e:
print(str(e))
pass
print("Ended extraction : " + extraction_type)
