def cond_percentile(self, df, cond):
cond = cond.fillna(0.0).astype(bool)
df, cond = self._align_bivariate(df, cond)
df = self._mask_non_index_member(df)
rank = rank_with_mask(df, mask=cond, axis=1, normalize=True)
return rank
def percentile(self, df, mask=None):
"""Return a DataFrame with values ranging from 0.0 to 1.0"""
df = self._align_univariate(df)
df = self._mask_non_index_member(df)
df = self._mask_df(df, mask)
rank = rank_with_mask(df, axis=1, normalize=True)
return rank
def percentile(self, df, mask=None):
"""Return a DataFrame with values ranging from 0.0 to 1.0"""
df = self._align_univariate(df)
df = self._mask_non_index_member(df)
df = self._mask_df(df, mask)
rank = rank_with_mask(df, axis=1, normalize=True)
return rank
def group_percentile(self, x, group):
x = self._align_univariate(x)
x = self._mask_non_index_member(x)
vals = np.unique(pd.Series(group.values.flatten()).dropna())
res = None
for val in vals:
mask = (group == val)
rank = rank_with_mask(x, mask=mask, axis=1, normalize=True)
if res is None:
res = rank
else:
res = res.fillna(rank)
return res
def group_rank(self, df, group, mask=None):
df = self._align_univariate(df)
df = self._mask_non_index_member(df)
df = self._mask_df(df, mask)
vals = np.unique(pd.Series(group.values.flatten()).dropna())
res = None
for val in vals:
mask = (group == val)
rank = rank_with_mask(df, mask=mask, axis=1, normalize=False)
if res is None:
res = rank
else:
res = res.fillna(rank)
return res
def group_percentile(self, df, group, mask=None):
df = self._align_univariate(df)
df = self._mask_non_index_member(df)
df = self._mask_df(df, mask)
vals = np.unique(pd.Series(group.values.flatten()).dropna())
res = None
for val in vals:
mask = (group == val)
rank = rank_with_mask(df, mask=mask, axis=1, normalize=True)
if res is None:
res = rank
else:
res = res.fillna(rank)
return res
def rank_standardize(factor_df, index_member=None):
"""
输入因子值, 将因子用排序分值重构，并处理到0-1之间(默认为升序——因子越大 排序分值越大(越好)
:param index_member:
:param factor_df: 因子值 (pandas.Dataframe类型),index为datetime, colunms为股票代码。
形如:
AAPL BA CMG DAL LULU
date
2016-06-24 0.165260 0.002198 0.085632 -0.078074 0.173832
2016-06-27 0.165537 0.003583 0.063299 -0.048674 0.180890
2016-06-28 0.135215 0.010403 0.059038 -0.034879 0.111691
2016-06-29 0.068774 0.019848 0.058476 -0.049971 0.042805
2016-06-30 0.039431 0.012271 0.037432 -0.027272 0.010902
:return: 排序重构后的因子值。 取值范围在0-1之间
"""
factor_df = jutil.fillinf(factor_df)
factor_df = _mask_non_index_member(factor_df, index_member)
return jutil.rank_with_mask(factor_df, axis=1, normalize=True)
def test_rank_percentile():
df = pd.DataFrame(np.random.rand(500, 3000))
res1 = jutil.rank_with_mask(df, axis=1, mask=None, normalize=False)
res2 = jutil.rank_with_mask(df, axis=1, mask=None, normalize=True)
print
def rank(self, df):
"""Return a DataFrame with values ranging from 0.0 to 1.0"""
df = self._align_univariate(df)
df = self._mask_non_index_member(df)
rank = rank_with_mask(df, axis=1, normalize=False)
return rank
