def __init__(self, model, x, y, cv_split=5, verbose=0, do_parallel=False):
"""
Parameters
-----------
model : scikit-learn supported model,
x : {array-like}, shape = [n_samples, n_features]
Training vectors, where n_samples is the number of samples
and n_features is the number of features.
y : {array-like}, shape = [n_samples]
Target Values
cv_split: int
Number of splits for cross_validation to calculate fitness.
verbose: 0 or 1
"""
self.model = model
self.n_features = x.shape[1]
self.toolbox = None
self.creator = self._create()
self.cv_split = cv_split
self.x = x
self.y = y
self.verbose = verbose
if self.verbose == 1:
print(
"Model {} will select best features among {} features using cv_split :{}."
.format(model, x.shape[1], cv_split))
print("Shape of train_x: {} and target: {}".format(
x.shape, y.shape))
self.do_parallel = do_parallel
self.final_fitness = []
self.fitness_in_generation = {}
self.best_ind = None
self.fit_obj = ff.FitenessFunction(self.cv_split)
def train(self):
fit_obj = ff.FitenessFunction(10)
feature_idx = np.where(np.asarray(self.best_ind) == 1)[0]
# print(feature_idx, self.best_ind)
fitness = fit_obj.calculate_fitness(self.model,
self.x.iloc[:,
feature_idx], self.y)
print("The accuracy using feature set {} is {}%".format(
feature_idx, fitness * 100))
def evaluate(self, individual):
fit_obj = ff.FitenessFunction()
np_ind = np.asarray(individual)
if np.sum(np_ind) == 0:
fitness = 0.0
else:
feature_idx = np.where(np_ind == 1)[0]
fitness = fit_obj.calculate_fitness\
(self.model,self.x[:,feature_idx], self.y, self.x_test[:,feature_idx],self.y_test,self.x_development[:,feature_idx],self.y_development)
if self.verbose == 1:
pass
return fitness,
def evaluate(self, individual):
fit_obj = ff.FitenessFunction(self.cv_split)
np_ind = np.asarray(individual)
if np.sum(np_ind) == 0:
fitness = 0.0
else:
feature_idx = np.where(np_ind == 1)[0]
fitness = fit_obj.calculate_fitness(self.model, self.x[:, feature_idx], self.y)
if self.verbose == 1:
print("Individual: {} Fitness_score: {} ".format(individual, fitness))
return fitness,
def evaluate(self, individual):
fit_obj = ff.FitenessFunction()
self.f1_score_test = fit_obj.get_f1()
np_ind = np.asarray(individual)
if np.sum(np_ind) == 0:
fitness = 0.0
else:
feature_idx = np.where(np_ind == 1)[0]
fitness = fit_obj.calculate_fitness\
(self.model,self.x[:,feature_idx], self.y, self.x_test[:,feature_idx],self.y_test,self.x_development[:,feature_idx],self.y_development)
if self.verbose == 1:
print("Individual: {} Fitness_score: {} ".format(
individual, fitness))
print(len(individual))
return fitness,
def f_per_particle(self, m, alpha=0.88):
fit_obj = ff.FitenessFunction()
# Get the subset of the features from the binary mask
if np.count_nonzero(m) == 0:
X_subset = self.X
else:
feature_idx = np.where(np.asarray(m) == 1)[0]
X_subset = self.X.iloc[:, feature_idx]
P, cv_set = fit_obj.calculate_fitness(self.classifier, X_subset, self.Y)
# Compute for the objective function
# j = (alpha * (1.0 - P)
# + (1.0 - alpha) * (1 - (X_subset.shape[1] / self.num_features)))
return (P,cv_set)