Merge pull request #156 from glevv/gastwirth-loc

Add Gastwirth's location estimator

README.md

@@ -15,6 +15,7 @@
 
 - Collection of measures of central tendency - `obscure_stats/central_tendency`:
     * Contraharmonic Mean;
+    * Gastwirth's Location;
     * Grenander's Mode;
     * Half-Sample Mode;
     * Hodges-Lehmann-Sen Location;

pyproject.toml

@@ -31,7 +31,7 @@ scipy = "^1.9.1"
 mypy = "^1.6.1"
 pytest = "^8.0.0"
 pytest-cov = "^5.0.0"
-ruff = "^0.7.0"
+ruff = "^0.8.0"
 hypothesis = "^6.103.1"
 hypothesis-pytest = "^0.19.0"
 

src/obscure_stats/association/association.py

@@ -468,7 +468,7 @@ def winsorized_correlation(x: np.ndarray, y: np.ndarray, k: float = 0.1) -> floa
         Input array.
     y : array_like
         Input array.
-    k : float
+    k : float, default = 0.1
         The percentages of values to winsorize on each side of the arrays.
 
     Returns

src/obscure_stats/central_tendency/__init__.py

@@ -2,6 +2,7 @@
 
 from .central_tendency import (
     contraharmonic_mean,
+    gastwirth_location,
     grenanders_m,
     half_sample_mode,
     hodges_lehmann_sen_location,
@@ -15,6 +16,7 @@
 
 __all__ = [
     "contraharmonic_mean",
+    "gastwirth_location",
     "grenanders_m",
     "half_sample_mode",
     "hodges_lehmann_sen_location",

src/obscure_stats/central_tendency/central_tendency.py

@@ -182,7 +182,7 @@ def standard_trimmed_harrell_davis_quantile(x: np.ndarray, q: float = 0.5) -> fl
     ----------
     x : array_like
         Input array.
-    q : float
+    q : float, default = 0.5
         Quantile value in range (0, 1).
 
     Returns
@@ -286,7 +286,7 @@ def tau_location(x: np.ndarray, c: float = 4.5) -> float:
     ----------
     x : array_like
         Input array.
-    c : float
+    c : float, default = 4.5
         Constant that filter outliers.
 
     Returns
@@ -310,7 +310,7 @@ def tau_location(x: np.ndarray, c: float = 4.5) -> float:
     return np.nansum(x * w) / np.nansum(w)
 
 
-def grenanders_m(x: np.ndarray, p: float = 1.5, k: int = 3) -> float:
+def grenanders_m(x: np.ndarray, p: float = 1.001, k: int = 2) -> float:
     """Calculate Grenander's Mode.
 
     This measure is a direct nonparametric estimation of the mode.
@@ -322,10 +322,10 @@ def grenanders_m(x: np.ndarray, p: float = 1.5, k: int = 3) -> float:
     ----------
     x : array_like
         Input array.
-    p : float
+    p : float, default = 1.001
         Smoothing constant.
-    k : int
-        The number of samples to exclude from the calculation.
+    k : int, default = 2
+        The number of samples to filter.
 
     Returns
     -------
@@ -339,6 +339,7 @@ def grenanders_m(x: np.ndarray, p: float = 1.5, k: int = 3) -> float:
     Annals of Mathematical Statistics, 36, 131-138.
     """
     x_sort = np.sort(x)
+    x_sort = x_sort.astype("float")
     x_sort = x_sort[np.isfinite(x_sort)]
 
     if p <= 1:
@@ -351,8 +352,41 @@ def grenanders_m(x: np.ndarray, p: float = 1.5, k: int = 3) -> float:
     if len(x_sort) <= k:
         return np.nan
 
+    # pre calculate diffs
+    diff = x_sort[k:] - x_sort[:-k]
+    # if the diffs are constant - return the value
+    if diff.sum() == 0.0:
+        return x_sort[0]
+    # to avoid division by zero
+    diff[diff == 0.0] = np.nan
+
     return (
         0.5
-        * np.sum((x_sort[k:] + x_sort[:-k]) / np.power(x_sort[k:] - x_sort[:-k], p))
-        / np.sum(np.power(x_sort[k:] - x_sort[:-k], -p))
+        * np.nansum((x_sort[k:] + x_sort[:-k]) / np.power(diff, p))
+        / np.nansum(np.power(diff, -p))
     )
+
+
+def gastwirth_location(x: np.ndarray) -> float:
+    """Calculate Gastwirth's location estimator.
+
+    This measure is more robust then average.
+
+    Parameters
+    ----------
+    x : array_like
+        Input array.
+
+    Returns
+    -------
+    gle : float
+        The value of the Gastwirth's location.
+
+    References
+    ----------
+    Gastwirth, J. L. (1966).
+    On Robust Procedures.
+    J. Amer. Statist. Assn., Vol. 61, pp. 929-948.
+    """
+    p33, p50, p66 = np.nanquantile(x, [1 / 3, 0.5, 2 / 3])
+    return 0.3 * p33 + 0.4 * p50 + 0.3 * p66

src/obscure_stats/skewness/skewness.py

@@ -412,8 +412,8 @@ def left_quantile_weight(x: np.ndarray, q: float = 0.25) -> float:
     ----------
     x : array_like
         Input array.
-    q : float
-        Quantile to use for the anchor.
+    q : float, default = 0.25
+        Quantile to use for the calculation, (0.0, 0.5)
 
     Returns
     -------
@@ -448,8 +448,8 @@ def right_quantile_weight(x: np.ndarray, q: float = 0.75) -> float:
     ----------
     x : array_like
         Input array.
-    q : float
-        Quantile to use for the anchor.
+    q : float, default = 0.75
+        Quantile to use for the calculation, (0.5, 1.0).
 
     Returns
     -------

src/obscure_stats/variation/variation.py

@@ -209,7 +209,7 @@ def renyi_entropy(x: np.ndarray, alpha: float = 2) -> float:
     ----------
     x : array_like
         Input array.
-    alpha : float
+    alpha : float, default = 2
         Order of the Rényi entropy
 
     Returns

tests/test_central_tendency.py

@@ -13,6 +13,7 @@
 
 from obscure_stats.central_tendency import (
     contraharmonic_mean,
+    gastwirth_location,
     grenanders_m,
     half_sample_mode,
     hodges_lehmann_sen_location,
@@ -26,6 +27,7 @@
 
 all_functions = [
     contraharmonic_mean,
+    gastwirth_location,
     grenanders_m,
     half_sample_mode,
     hodges_lehmann_sen_location,