def fit(self, X, y):
X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = self.test_size, random_state = self.random_state)
dim = X.shape[1]
self.indices_ = tuple(range(dim))
self.subsets_ = [self.indices_]
score = self._cal_score(X_train, X_test, y_train, y_test,self.indices_)
self.scores_ = [score]
while dim > self.k_features:
scores = []
subsets = []
for p in combinations(self.indices_, r = dim -1):
score = self._cal_score(X_train, X_test, y_train, y_test,p)
scores.append(score)
subsets.append(p)
best = np.argmax(scores)
self.indices_ = subsets[best]
self.subsets_.append(self.indices_)
dim -= 1
self.scores_.append(scores[best])
self.k_score_ = self.scores_[-1]
return self
def model_auto_tpot(
df,
colX, coly,
outfolder="aaserialize/",
model_type="regressor/classifier",
train_size=0.5,
generation=1,
population_size=5,
verbosity=2,
):
""" Automatic training of Xmat--->Y, Generate SKlearn code in outfile
Very Slow Process, use lower number of Sample
:param Xmat:
:param y:
:param outfolder:
:param model_type:
:param train_size:
:param generation:
:param population_size:
:param verbosity:
:return:
"""
tpot = import_("tpot")
X = df[colX].values
y = df[coly].values
X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.5, test_size=0.5)
if model_type == "regressor":
clf = tpot.TPOTRegressor(
generations=generation, population_size=population_size, verbosity=verbosity
)
elif model_type == "classifier":
clf = tpot.TPOTClassifier(
generations=generation, population_size=population_size, verbosity=verbosity
)
print("Start")
clf.fit(X_train, y_train)
score = tpot.score(X_test, y_test)
print("score", score)
file1 = outfolder + "/tpot_regression_pipeline_" + str(np.random.randint(1000, 9999)) + ".py"
tpot.export(file1)
return file1
X = name_letters
X = np.array(X).reshapre(-1, 1)
y = np.where(labeled_names[ind, 1] == 'male', 0, 1)
from sklearn import preprocessing
lb = preprocessing.LabelBinarizer()
lb.fit(X)
X2 = lb.transform(X)
from sklearn.preprocessing import train_test_split
X_train2, y_train2, X_test2, y_test2 = train_test_split(X2,
y,
test_size=0.4,
random_state=42)
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.4,
random_state=42)
from sklearn.naive_bayes import MultinomialNB
from sklearn.metrics import accuracy_score, confusion_matrix
clf = MultinomialNB.new(alpha=0.1, fit_prior=True)
clf.fit(X_train2, y_train2)
y_train_pred = clf.predict(x_train2)
y_test_pred = clf.predict(x_test2)
def train_val_test(df, y):
train, test1 = train_test_split(df, test_size=.70, stratify=y, random_state=42)
val, test = train_test_split(test1, test_size=.50, stratify=y, random_stat=42)
return train, val, test
def train_test(self, test_size):
self.x_train, self.x_test, self.y_train, self.y_test = train_test_split(
self.x, self.y, test_size=test_size, random_state=0)
X = dataset.ilocs[:, :-1].values
y = dataset.ilocs[:, :4].values
#Encoding Categorical Data
from sklearn import LabelEncoder, OneHotEncoder
labelencoder = LabelEncoder()
X[:, 3] = labelencoder.fit_transform(X[:, 3])
onehotencoder = OneHotEncoder(categorical_features=[3])
X = onehotencoder.fit_transform(X).toarray()
#exclure l'index 0
X = X[:, 1:]
#Split Data Train and DataSet
from sklearn.preprocessing import train_test_split
X_train, y_train, X_test, y_test = train_test_split(X,
y,
test_size=1 / 3,
random_state=0)
from sklearn.Preprocessing import train_test_split
X_train, y_train, X_test, y_test = train_test_split(X,
y,
test_seize=1 / 3,
random_state=0)
# Importing the libraries
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
# Importing the dataset
dataset = pd.read_csv('50_Startups.csv')
r_matr = dataset.iloc[:, -1].values # Fill in missing data with mean method from sklearn.preprocessing import Imputer imputer = Imputer(missing_values="NaN", strategy="mean", axis=0) imputer = imputer.fit(f_matr[:, [1,2]]) f_matr[:, [1,2]] = imputer.transform(f_matr[:, [1,2]]) # Categorial Data splitting from sklearn.preprocessing import LabelEncoder, OneHotEncoder country_encoder = LabelEncoder() f_matr[:, 0] = country_encoder.fit_transform(f_matr[:, 0]) onehotencoder = OneHotEncoder(categorical_features=[0]) f_matr = onehotencoder.fit_transform(features_matrix).toarray() response_encoder = LabelEncoder() r_matr = response_encoder.fit_transform(r_matr) # Separating training and testing set from sklearn.preprocessing import train_test_split f_matr_train, f_matr_test,\ r_matr_train, r_matr_test = train_test_split(f_matr, r_matr, test_size=0.2, random_state=0) # Feature Scaling -- age and salary columns should be same range from sklearn.preprocessing import StandardScaler standard_scaler = StandardScaler() f_matr_train = standard_scaler.fit_transform(f_matr_train) f_matr_test = standard_scaler.fit_transform(f_matr_test)