def __init__(self,
data=None,
id=None,
time=None,
name='panel data model',
formula=None,
var_transform=False,
model='Fixed',
*args,
**kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建公式
self._formula = formula
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [
(self._copy_data.columns[i], '_'.join(['var', str(i)]))
for i in range(len(self._copy_data.columns))
]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([
('_'.join(['var', str(i)]), self._copy_data.columns[i])
for i in range(len(self._copy_data.columns))
])
self._generated_variables = [
item[1] for item in self._variables_mapping
]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key, self._variables_mapping_dict[key],
self._formula)
#self._formula = Formula(self._formula)
self._model = model
# 导入plm库
self._renv._R.r('library("plm")')
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
# 把数据格式转换为面板格式
fmt_str = 'pdata <- pdata.frame(mdata, index=c("{0}","{1}"))'
self._renv._R.r(fmt_str.format(id, time))
self._summary = importr('base').summary
def __init__(self,
data=None,
name='Propensity Score Matching',
formula=None,
method='nearest',
distance='logit',
var_transform=False,
*args,
**kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建matchit
self._formula = formula
# 创建方法
self._method = method
# 创建距离
self._distance = distance
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [
(self._copy_data.columns[i], '_'.join(['var', str(i)]))
for i in range(len(self._copy_data.columns))
]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([
('_'.join(['var', str(i)]), self._copy_data.columns[i])
for i in range(len(self._copy_data.columns))
])
self._generated_variables = [
item[1] for item in self._variables_mapping
]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key, self._variables_mapping_dict[key],
self._formula)
# self._formula = Formula(self._formula)
# 导入MatchIt库
self._renv._R.r('library("MatchIt")')
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __init__(self, data=None, name='Propensity Score Matching', formula=None, method='nearest',
distance = 'logit', var_transform=False, *args, **kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建matchit
self._formula = formula
# 创建方法
self._method = method
# 创建距离
self._distance = distance
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [(self._copy_data.columns[i],'_'.join(['var',str(i)]))
for i in range(len(self._copy_data.columns))]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([('_'.join(['var',str(i)]),self._copy_data.columns[i]) for i in range(len(self._copy_data.columns))])
self._generated_variables = [item[1] for item in self._variables_mapping]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key,self._variables_mapping_dict[key],self._formula)
# self._formula = Formula(self._formula)
# 导入MatchIt库
self._renv._R.r('library("MatchIt")')
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __init__(self, data=None, name='regression', formula=None, var_transform=False, family='binomial(link=logit)', *args, **kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建公式
self._formula = formula
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
self._family = family
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [(self._copy_data.columns[i],'_'.join(['var',str(i)]))
for i in range(len(self._copy_data.columns))]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([('_'.join(['var',str(i)]),self._copy_data.columns[i]) for i in range(len(self._copy_data.columns))])
self._generated_variables = [item[1] for item in self._variables_mapping]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key,self._variables_mapping_dict[key],self._formula)
# self._formula = Formula(self._formula)
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __init__(self, data=None, name='regression', formula=None, var_transform=False, *args, **kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建公式
self._formula = formula
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [(self._copy_data.columns[i],'_'.join(['var',str(i)]))
for i in range(len(self._copy_data.columns))]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([('_'.join(['var',str(i)]),self._copy_data.columns[i]) for i in range(len(self._copy_data.columns))])
self._generated_variables = [item[1] for item in self._variables_mapping]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key,self._variables_mapping_dict[key],self._formula)
self._formula = Formula(self._formula)
self._lm = importr('stats').lm
self._summary = importr('base').summary
def __call__(self):
if self._copy_data is not None:
lm_obj = self._lm(self._formula, data=REnv.python_to_r_object(self._copy_data))
self._result['lm'] = self._renv[lm_obj]
self._result['summary'] = self._renv[self._summary(lm_obj)]
else:
print('Data is not defined!')
raise Exception
return self._result
def __init__(self, data=None, func=None):
self._Renv = REnv()
# 导入R中的boot库
self._Renv._R.r('library("boot")')
# 定义函数
self._func = '{0} <- {1}'.format('myfunc',func)
self._Renv._R.r(self._func)
# 导入数据
self._Renv['data'] = REnv.python_to_r_object(data)
class PropensityScoreMatching(BasicModel):
def __init__(self, data=None, name='Propensity Score Matching', formula=None, method='nearest',
distance = 'logit', var_transform=False, *args, **kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建matchit
self._formula = formula
# 创建方法
self._method = method
# 创建距离
self._distance = distance
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [(self._copy_data.columns[i],'_'.join(['var',str(i)]))
for i in range(len(self._copy_data.columns))]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([('_'.join(['var',str(i)]),self._copy_data.columns[i]) for i in range(len(self._copy_data.columns))])
self._generated_variables = [item[1] for item in self._variables_mapping]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key,self._variables_mapping_dict[key],self._formula)
# self._formula = Formula(self._formula)
# 导入MatchIt库
self._renv._R.r('library("MatchIt")')
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __call__(self):
if self._copy_data is not None:
# estimate propensity scores and create matched data set using 'matchit'
matching_it_str = 'match_it <- matchit({0}, data = mdata, method = "{1}", distance = "{2}")'.format(self._formula, self._method, self._distance)
self._renv._R.r(matching_it_str)
self._renv._R.r('psm <- match_it$nn')
self._result['psm'] = self._renv['psm']
self._renv._R.r('matched_data <- match.data(match_it,distance ="pscore")')
self._result['matched_data'] = self._renv['matched_data']
else:
print('Data is not defined!')
raise Exception
return self._result
def __repr__(self):
dot_line = '-'*80
title = 'Propensity Score Matching: {}\n'.format(self._origin_formula)
return ''.join([title,
dot_line,'\n',
self._result['psm'].__repr__(),'\n',
dot_line,'\n'])
@property
def matched_data(self):
return self._result['matched_data']
class Logistic(BasicModel):
def __init__(self, data=None, name='regression', formula=None, var_transform=False, family='binomial(link=logit)', *args, **kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建公式
self._formula = formula
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
self._family = family
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [(self._copy_data.columns[i],'_'.join(['var',str(i)]))
for i in range(len(self._copy_data.columns))]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([('_'.join(['var',str(i)]),self._copy_data.columns[i]) for i in range(len(self._copy_data.columns))])
self._generated_variables = [item[1] for item in self._variables_mapping]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key,self._variables_mapping_dict[key],self._formula)
# self._formula = Formula(self._formula)
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __call__(self):
if self._copy_data is not None:
glm_str = 'glm({0},family={1},data=mdata)'
print(glm_str.format(self._formula, self._family))
glm_obj = self._renv._R.r(glm_str.format(self._formula, self._family))
self._result['lm'] = self._renv[glm_obj]
self._result['summary'] = self._renv[self._summary(glm_obj)]
else:
print('Data is not defined!')
raise Exception
return self._result
def __repr__(self):
dot_line = '-'*80
title = 'Binary Response Regression Result: {}\n'.format(self._origin_formula)
nobs = 'Number of Observation: {}'.format(self._data.shape[0])
return ''.join([title,
nobs,'\n',
dot_line,'\n',
self.coefs.__repr__(),'\n',
dot_line,'\n'])
@property
def coefs(self):
coefs = self._result['summary']['coefficients']
if self._var_transform:
indexes = coefs.index
new_indexes = []
for ind in indexes:
if ind in self._variables_mapping_dict_reversed:
new_indexes.append(self._variables_mapping_dict_reversed[ind])
elif ind == '(Intercept)':
new_indexes.append('常数')
else:
new_indexes.append(ind)
coefs.index = new_indexes
return coefs
class PropensityScoreMatching(BasicModel):
def __init__(self,
data=None,
name='Propensity Score Matching',
formula=None,
method='nearest',
distance='logit',
var_transform=False,
*args,
**kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建matchit
self._formula = formula
# 创建方法
self._method = method
# 创建距离
self._distance = distance
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [
(self._copy_data.columns[i], '_'.join(['var', str(i)]))
for i in range(len(self._copy_data.columns))
]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([
('_'.join(['var', str(i)]), self._copy_data.columns[i])
for i in range(len(self._copy_data.columns))
])
self._generated_variables = [
item[1] for item in self._variables_mapping
]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key, self._variables_mapping_dict[key],
self._formula)
# self._formula = Formula(self._formula)
# 导入MatchIt库
self._renv._R.r('library("MatchIt")')
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __call__(self):
if self._copy_data is not None:
# estimate propensity scores and create matched data set using 'matchit'
matching_it_str = 'match_it <- matchit({0}, data = mdata, method = "{1}", distance = "{2}")'.format(
self._formula, self._method, self._distance)
self._renv._R.r(matching_it_str)
self._renv._R.r('psm <- match_it$nn')
self._result['psm'] = self._renv['psm']
self._renv._R.r(
'matched_data <- match.data(match_it,distance ="pscore")')
self._result['matched_data'] = self._renv['matched_data']
else:
print('Data is not defined!')
raise Exception
return self._result
def __repr__(self):
dot_line = '-' * 80
title = 'Propensity Score Matching: {}\n'.format(self._origin_formula)
return ''.join([
title, dot_line, '\n', self._result['psm'].__repr__(), '\n',
dot_line, '\n'
])
@property
def matched_data(self):
return self._result['matched_data']
class Logistic(BasicModel):
def __init__(self,
data=None,
name='regression',
formula=None,
var_transform=False,
family='binomial(link=logit)',
*args,
**kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建公式
self._formula = formula
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
self._family = family
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [
(self._copy_data.columns[i], '_'.join(['var', str(i)]))
for i in range(len(self._copy_data.columns))
]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([
('_'.join(['var', str(i)]), self._copy_data.columns[i])
for i in range(len(self._copy_data.columns))
])
self._generated_variables = [
item[1] for item in self._variables_mapping
]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key, self._variables_mapping_dict[key],
self._formula)
# self._formula = Formula(self._formula)
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
self._summary = importr('base').summary
def __call__(self):
if self._copy_data is not None:
glm_str = 'glm({0},family={1},data=mdata)'
print(glm_str.format(self._formula, self._family))
glm_obj = self._renv._R.r(
glm_str.format(self._formula, self._family))
self._result['lm'] = self._renv[glm_obj]
self._result['summary'] = self._renv[self._summary(glm_obj)]
else:
print('Data is not defined!')
raise Exception
return self._result
def __repr__(self):
dot_line = '-' * 80
title = 'Binary Response Regression Result: {}\n'.format(
self._origin_formula)
nobs = 'Number of Observation: {}'.format(self._data.shape[0])
return ''.join([
title, nobs, '\n', dot_line, '\n',
self.coefs.__repr__(), '\n', dot_line, '\n'
])
@property
def coefs(self):
coefs = self._result['summary']['coefficients']
if self._var_transform:
indexes = coefs.index
new_indexes = []
for ind in indexes:
if ind in self._variables_mapping_dict_reversed:
new_indexes.append(
self._variables_mapping_dict_reversed[ind])
elif ind == '(Intercept)':
new_indexes.append('常数')
else:
new_indexes.append(ind)
coefs.index = new_indexes
return coefs
class Plm(BasicModel):
def __init__(self,
data=None,
id=None,
time=None,
name='panel data model',
formula=None,
var_transform=False,
model='Fixed',
*args,
**kwargs):
super().__init__(data=data, name=name, *args, **kwargs)
# 创建REnv实例
self._renv = REnv()
# 创建公式
self._formula = formula
# 原始公式
self._origin_formula = self._formula
self._copy_data = deepcopy(data)
# 转换变量，特别是那些变量是中文的
self._var_transform = var_transform
if self._var_transform:
self._variables_mapping = [
(self._copy_data.columns[i], '_'.join(['var', str(i)]))
for i in range(len(self._copy_data.columns))
]
self._variables_mapping_dict = dict(self._variables_mapping)
self._variables_mapping_dict_reversed = dict([
('_'.join(['var', str(i)]), self._copy_data.columns[i])
for i in range(len(self._copy_data.columns))
])
self._generated_variables = [
item[1] for item in self._variables_mapping
]
self._copy_data.columns = self._generated_variables
for key in self._variables_mapping_dict:
self._formula = re.sub(key, self._variables_mapping_dict[key],
self._formula)
#self._formula = Formula(self._formula)
self._model = model
# 导入plm库
self._renv._R.r('library("plm")')
# 把数据转换为R数据格式
self._renv['mdata'] = self._renv.python_to_r_object(self._copy_data)
# 把数据格式转换为面板格式
fmt_str = 'pdata <- pdata.frame(mdata, index=c("{0}","{1}"))'
self._renv._R.r(fmt_str.format(id, time))
self._summary = importr('base').summary
def __call__(self):
if self._copy_data is not None:
plm_str = 'plm({0},data=pdata,model="{1}")'
print(plm_str.format(self._formula, self._model))
plm_obj = self._renv._R.r(
plm_str.format(self._formula, self._model))
self._result['plm'] = self._renv[plm_obj]
self._result['summary'] = self._renv[self._summary(plm_obj)]
else:
print('Data is not defined!')
raise Exception
return self._result
def __repr__(self):
dot_line = '-' * 80
title = 'Panel Data Result: {}\n'.format(self._origin_formula)
nobs = 'Number of Observation: {}'.format(self._data.shape[0])
return ''.join([
title, nobs, '\n', dot_line, '\n',
self.coefs.__repr__(), '\n', dot_line, '\n'
])
@property
def coefs(self):
coefs = self._result['summary']['coefficients']
if self._var_transform:
indexes = coefs.index
new_indexes = []
for ind in indexes:
if ind in self._variables_mapping_dict_reversed:
new_indexes.append(
self._variables_mapping_dict_reversed[ind])
elif ind == '(Intercept)':
new_indexes.append('常数')
else:
new_indexes.append(ind)
coefs.index = new_indexes
return coefs