Uploaded Test files

venv/Lib/site-packages/sklearn/utils/stats.py

@@ -0,0 +1,18 @@
+import numpy as np
+
+from .extmath import stable_cumsum
+
+
+def _weighted_percentile(array, sample_weight, percentile=50):
+    """
+    Compute the weighted ``percentile`` of ``array`` with ``sample_weight``.
+    """
+    sorted_idx = np.argsort(array)
+
+    # Find index of median prediction for each sample
+    weight_cdf = stable_cumsum(sample_weight[sorted_idx])
+    percentile_idx = np.searchsorted(
+        weight_cdf, (percentile / 100.) * weight_cdf[-1])
+    # in rare cases, percentile_idx equals to len(sorted_idx)
+    percentile_idx = np.clip(percentile_idx, 0, len(sorted_idx)-1)
+    return array[sorted_idx[percentile_idx]]