<p>滞后0的PACF始终为1(参见例<a href="https://www.itl.nist.gov/div898/handbook/pmc/section4/pmc4463.htm" rel="nofollow noreferrer">here</a>),因此其置信区间为[1,1]</p>
<p>计算CI的<a href="https://github.com/statsmodels/statsmodels/blob/0551e89f7b4bae6cf6a5711b29e8112a9edc8cde/statsmodels/tsa/stattools.py#L928-L931" rel="nofollow noreferrer">the code snippet</a>的最后一行确保了这一点:</p>
<pre><code>varacf = 1. / len(x) # for all lags >=1
interval = stats.norm.ppf(1. - alpha / 2.) * np.sqrt(varacf)
confint = np.array(lzip(ret - interval, ret + interval))
confint[0] = ret[0] # fix confidence interval for lag 0 to varpacf=0
</code></pre>
<p>(另见<a href="https://github.com/statsmodels/statsmodels/issues/1969" rel="nofollow noreferrer">issue 1969</a>,其中这是固定的)</p>
<p>由于对0滞后不感兴趣,您通常使PACF绘图从滞后1开始(如R的<a href="https://www.rdocumentation.org/packages/tseries/versions/0.1-2/topics/pacf" rel="nofollow noreferrer">pacf function</a>)。这可以通过<code>zero=False</code>实现:</p>
<pre><code>sm.graphics.tsa.plot_pacf(x, ax=axes[0], zero=True, title='zero=True (default)')
sm.graphics.tsa.plot_pacf(x, ax=axes[1], zero=False, title='zero=False')
</code></pre>
<p><a href="https://i.stack.imgur.com/uhyFb.png" rel="nofollow noreferrer"><img src="https://i.stack.imgur.com/uhyFb.png" alt="enter image description here"/></a></p>