class MergedNB(NaiveBayes):
def __init__(self, whether_discrete):
NaiveBayes.__init__(self)
self._whether_discrete = np.array(whether_discrete)
self._whether_continuous = ~self._whether_discrete
self._multinomial, self._gaussian = MultinomialNB(), GaussianNB()
def feed_data(self, x, y, sample_weights=None):
x = np.array(x)
self._multinomial.feed_data(x[:, self._whether_discrete], y,
sample_weights)
y = self._multinomial["y"]
self.label_dic = self._multinomial.label_dic
self._cat_counter = self._multinomial["cat_counter"]
self._gaussian.feed_data(x[:, self._whether_continuous], y,
sample_weights)
self._gaussian.label_dic = self._multinomial.label_dic
def feed_sample_weights(self, sample_weights=None):
self._multinomial.feed_sample_weights(sample_weights)
self._gaussian.feed_sample_weights(sample_weights)
def _fit(self, lb):
self._multinomial.fit()
self._gaussian.fit()
discrete_func, continuous_func = self._multinomial[
"func"], self._gaussian["func"]
def func(input_x, tar_category):
input_x = np.array(input_x)
return discrete_func(
input_x[self._whether_discrete].astype(np.int),
tar_category) * continuous_func(
input_x[self._whether_continuous], tar_category)
return func
def _transfer_x(self, x):
_feat_dics = self._multinomial["feat_dics"]
idx = 0
for d, discrete in enumerate(self._whether_discrete):
if not discrete:
x[d] = float(x[d])
else:
x[d] = _feat_dics[idx][x[d]]
if discrete:
idx += 1
return x
if __name__ == '__main__':
import time
xs, ys = DataUtil.get_dataset("mushroom", "../../_Data/mushroom.txt", tar_idx=0)
nb = MultinomialNB()
nb.feed_data(xs, ys)
xs, ys = nb["x"].tolist(), nb["y"].tolist()
train_num = 6000
x_train, x_test = xs[:train_num], xs[train_num:]
y_train, y_test = ys[:train_num], ys[train_num:]
learning_time = time.time()
nb = GaussianNB()
nb.fit(x_train, y_train)
learning_time = time.time() - learning_time
estimation_time = time.time()
nb.evaluate(x_train, y_train)
nb.evaluate(x_test, y_test)
estimation_time = time.time() - estimation_time
print(
"Model building : {:12.6} s\n"
"Estimation : {:12.6} s\n"
"Total : {:12.6} s".format(
learning_time, estimation_time,
learning_time + estimation_time
)
)
class MergedNB(NaiveBayes):
MergedNBTiming = Timing()
def __init__(self, **kwargs):
super(MergedNB, self).__init__(**kwargs)
self._multinomial, self._gaussian = MultinomialNB(), GaussianNB()
wc = kwargs.get("whether_continuous")
if wc is None:
self._whether_discrete = self._whether_continuous = None
else:
self._whether_continuous = np.asarray(wc)
self._whether_discrete = ~self._whether_continuous
@MergedNBTiming.timeit(level=1, prefix="[API] ")
def feed_data(self, x, y, sample_weight=None):
if sample_weight is not None:
sample_weight = np.asarray(sample_weight)
x, y, wc, features, feat_dics, label_dic = DataUtil.quantize_data(
x, y, wc=self._whether_continuous, separate=True)
if self._whether_continuous is None:
self._whether_continuous = wc
self._whether_discrete = ~self._whether_continuous
self.label_dic = label_dic
discrete_x, continuous_x = x
cat_counter = np.bincount(y)
self._cat_counter = cat_counter
labels = [y == value for value in range(len(cat_counter))]
labelled_x = [discrete_x[ci].T for ci in labels]
self._multinomial._x, self._multinomial._y = x, y
self._multinomial._labelled_x, self._multinomial._label_zip = labelled_x, list(zip(labels, labelled_x))
self._multinomial._cat_counter = cat_counter
self._multinomial._feat_dics = [dic for i, dic in enumerate(feat_dics) if self._whether_discrete[i]]
self._multinomial._n_possibilities = [len(feats) for i, feats in enumerate(features)
if self._whether_discrete[i]]
self._multinomial.label_dic = label_dic
labelled_x = [continuous_x[label].T for label in labels]
self._gaussian._x, self._gaussian._y = continuous_x.T, y
self._gaussian._labelled_x, self._gaussian._label_zip = labelled_x, labels
self._gaussian._cat_counter, self._gaussian.label_dic = cat_counter, label_dic
self.feed_sample_weight(sample_weight)
@MergedNBTiming.timeit(level=1, prefix="[Core] ")
def feed_sample_weight(self, sample_weight=None):
self._multinomial.feed_sample_weight(sample_weight)
self._gaussian.feed_sample_weight(sample_weight)
@MergedNBTiming.timeit(level=1, prefix="[Core] ")
def _fit(self, lb):
self._multinomial.fit()
self._gaussian.fit()
p_category = self._multinomial.get_prior_probability(lb)
discrete_func, continuous_func = self._multinomial["func"], self._gaussian["func"]
def func(input_x, tar_category):
input_x = np.asarray(input_x)
return discrete_func(
input_x[self._whether_discrete].astype(np.int), tar_category) * continuous_func(
input_x[self._whether_continuous], tar_category) / p_category[tar_category]
return func
@MergedNBTiming.timeit(level=1, prefix="[Core] ")
def _transfer_x(self, x):
feat_dics = self._multinomial["feat_dics"]
idx = 0
for d, discrete in enumerate(self._whether_discrete):
if not discrete:
x[d] = float(x[d])
else:
x[d] = feat_dics[idx][x[d]]
if discrete:
idx += 1
return x
np.random.shuffle(_data)
train_num = 6000
xs = _data
ys = [xx.pop(0) for xx in xs]
nb = MultinomialNB()
nb.feed_data(xs, ys)
xs, ys = nb["x"].tolist(), nb["y"].tolist()
train_x, test_x = xs[:train_num], xs[train_num:]
train_y, test_y = ys[:train_num], ys[train_num:]
train_num = 6000
train_data = _data[:train_num]
test_data = _data[train_num:]
learning_time = time.time()
nb = GaussianNB()
nb.fit(train_x, train_y)
learning_time = time.time() - learning_time
estimation_time = time.time()
nb.estimate(train_x, train_y)
nb.estimate(test_x, test_y)
estimation_time = time.time() - estimation_time
print("Model building : {:12.6} s\n"
"Estimation : {:12.6} s\n"
"Total : {:12.6} s".format(
learning_time, estimation_time, learning_time + estimation_time))