Adding support for Series.autocorr

docs/cudf/source/api_docs/series.rst

@@ -118,6 +118,7 @@ Computations / descriptive stats
    Series.abs
    Series.all
    Series.any
+   Series.autocorr
    Series.ceil
    Series.clip
    Series.corr

python/cudf/cudf/core/frame.py

@@ -1919,7 +1919,6 @@ def round(self, decimals=0, how="half_even"):
         2   0.7   0.0
         3   0.2   0.0
         """
-
         if isinstance(decimals, cudf.Series):
             decimals = decimals.to_pandas()
 

python/cudf/cudf/core/series.py

@@ -2824,6 +2824,31 @@ def corr(self, other, method="pearson", min_periods=None):
 
         return lhs._column.corr(rhs._column)
 
+    def autocorr(self, lag=1):
+        """Compute the lag-N autocorrelation. This method computes the Pearson
+        correlation between the Series and its shifted self.
+
+        Parameters
+        ----------
+        lag : int, default 1
+            Number of lags to apply before performing autocorrelation.
+
+        Returns
+        -------
+        result : float
+            The Pearson correlation between self and self.shift(lag).
+
+        Examples
+        --------
+        >>> import cudf
+        >>> s = cudf.Series([0.25, 0.5, 0.2, -0.05])
+        >>> s.autocorr()
+        0.10355263309024071
+        >>> s.autocorr(lag=2)
+        -0.9999999999999999
+        """
+        return self.corr(self.shift(lag))
+
     def isin(self, values):
         """Check whether values are contained in Series.
 

python/cudf/cudf/tests/test_series.py

@@ -1348,6 +1348,23 @@ def test_nullable_bool_dtype_series(data, bool_dtype):
     assert_eq(psr, gsr.to_pandas(nullable=True))
 
 
+@pytest.mark.parametrize(
+    "cudf_series",
+    [
+        cudf.Series([0.25, 0.5, 0.2, -0.05]),
+        cudf.Series([0, 1, 2, np.nan, 4, cudf.NA, 6]),
+    ],
+)
+@pytest.mark.parametrize("lag", [1, 2, 3, 4])
+def test_autocorr(cudf_series, lag):
+    psr = cudf_series.to_pandas()
+
+    cudf_corr = cudf_series.autocorr(lag=lag)
+    pd_corr = psr.autocorr(lag=lag)
+
+    assert_eq(pd_corr, cudf_corr)
+
+
 @pytest.mark.parametrize(
     "data",
     [