__author__?=?‘Thesharing‘

import?math
import?random
import?numpy?as?np
import?pandas
from?sklearn.base?import?clone?ClassifierMixin?RegressorMixin
from?sklearn.tree?import?DecisionTreeClassifier
from?sklearn.tree?import?DecisionTreeRegressor
from?sklearn.ensemble?import?AdaBoostRegressor
from?sklearn.metrics?import?accuracy_score


class?MultiBoostClassifier:
????def?__init__（self?base_estimator=DecisionTreeClassifier（）?n_estimators=10）:

????????pandas.options.mode.chained_assignment?=?None

????????self.base_estimator?=?base_estimator
????????self.n_estimators?=?n_estimators
????????self.beta?=?np.empty（0）
????????self.error_rate?=?np.empty（0）
????????self.data_size?=?0
????????self.y_range?=?np.empty（0）
????????self._estimators?=?np.empty（0）
????????self._class_num?=?0

????def?fit（self?X:?pandas.Dataframe?y）:
????????#?Initialization
????????self._estimators?=?np.empty（self.n_estimators?dtype=ClassifierMixin）
????????self.beta?=?np.empty（[self.n_estimators]）
????????self.error_rate?=?np.empty（[self.n_estimators]）
????????self.data_size?=?X.shape[0]
????????if?isinstance（y?pandas.Dataframe）:
????????????y?=?y.values.reshape（（-1））
????????self._get_y_dict（y）
????????y?=?self._transform_y（y）
????????self.y_range?=?np.unique（y）
????????#?S‘?=?S?with?instance?weights?assigned?to?be?1.
????????sample_X?sample_y?sample_weight?=?self._reset_sample_and_standardize（X?y?mode=‘uniform‘）
????????#?set?k?=?1
????????k?=?1
????????#?for?t?=?1?to?T?{
????????for?t?in?range（self.n_estimators）:
????????????#?if?Ik?=?t?then
????????????if?self._get_iteration_number（k）?==?t:
????????????????#?reset?S‘?to?random?weights?drawn?from?the?continuous?Poisson?distribution.
????????????????#?standardize?S‘?to?sum?to?n.
????????????????sample_X?sample_y?sample_weight?=?self._reset_sample_and_standardize（X?y）
????????????????#?increment?k
????????????????k?=?k?+?1
????????????#?Ct?=?baseLearn（S‘）
????????????temp?=?clone（self.base_estimator）.fit（sample_X?sample_y?sample_weight）
????????????self._estimators[t]?=?temp
????????????estimate_y?=?temp.predict（sample_X）
????????????self.error_rate[t]?=?1?-?accuracy_score（sample_y?estimate_y?sample_weight=sample_weight）
????????????while?self.error_rate[t]?>?1?-?1.?/?self._class_num:
????????????????#?reset?S‘?to?random?weights?drawn?from?the?continuous?Poisson?distribution.
????????????????#?standardize?S‘?to?sum?to?n.
????????????????sample_X?sample_y?sample_weight?=?self._reset_sample_and_standardize（X?y）
????????????????#?increment?k
????????????????k?=?k?+?1
????????????????#?go?to?step?8
????????????????temp.fit（sample_X?sample_y?sample_weight）
????????????????self._estimators[t]?=?temp
????????????????estimate_y?=?temp.predict（sample_X）
????????????????self.error_rate[t]?=?1?-?accuracy_score（sample_y?estimate