DeterministicProcess 给我“‘NoneType’对象没有属性‘rule_code’”

Question

所以我试着运行这段代码：

from statsmodels.tsa.deterministic import CalendarFourier

y1 = close.copy()
y1.index = pd.to_datetime(y1.index)

fourier = CalendarFourier(freq='M',order=4)
dp1 = DeterministicProcess(constant=True,
                           index=y1.index,
                           order=1,
                           seasonal=True,
                           drop=True)
                           # additional_terms=[fourier])
X_time = dp1.in_sample()
X_time['NewYearsDay'] = (X_time.index.dayofyear == 1)

model = LinearRegression(fit_intercept=False)
model.fit(X_time,y1)
y_deseason = y1 - model.predict(X_time)
y_deseason.name = 'sales_deseasoned'

ax = y_deseason.plot()
ax.set_title('Sales of School and Office Supplies (deseasonalized)');

并导致：

AttributeError                            Traceback (most recent call last)
Input In [136], in <cell line: 7>()
      4 y1.index = pd.to_datetime(y1.index)
      6 fourier = CalendarFourier(freq='M',order=4)
----> 7 dp1 = DeterministicProcess(constant=True,
      8                            index=y1.index,
      9                            order=1,
     10                            seasonal=True,
     11                            drop=True)
     12                            # additional_terms=[fourier])
     13 X_time = dp1.in_sample()

File ~\Anaconda3\lib\site-packages\statsmodels\tsa\deterministic.py:1179, in DeterministicProcess.__init__(self, index, period, constant, order, seasonal, fourier, additional_terms, drop)
   1177 if (seasonal or fourier) and period is None:
   1178     if period is None:
-> 1179         self._period = period = freq_to_period(self._index_freq)
   1180 if seasonal:
   1181     period = required_int_like(period, "period")

File ~\Anaconda3\lib\site-packages\statsmodels\tsa\tsatools.py:796, in freq_to_period(freq)
    794 if not isinstance(freq, offsets.DateOffset):
    795     freq = to_offset(freq)  # go ahead and standardize
--> 796 freq = freq.rule_code.upper()
    798 if freq == "A" or freq.startswith(("A-", "AS-")):
    799     return 1

AttributeError: 'NoneType' object has no attribute 'rule_code'

我想可能是因为datetime类型曾经是“object”，所以又改成了“datetime”，但还是报错。 知道这里出了什么问题吗？ 高度赞赏

Answer 1

 dp1 = DeterministicProcess(constant=True,  index=y1.index,
  order=1,
 seasonal=True, ## Change this to False
 drop=True,
 additional_terms=[fourier])

使这个seasonal=False应该可以帮助您提供输出...原因可能在您正在处理的数据中...如果季节性有重复数据（具有多个值的相同日期或重复的索引将不起作用..所以季节性false 使其工作或通过聚合或删除重复项（keep='first'）修复重复日期。

要修复第二个选项是在创建索引时，我们需要使用 resample (pandas) 1 天，然后使用seasonal = True为我工作。

y1 = y1.set_index("date").resample('1D').pad()