diff --git a/neighbours/distance.py b/neighbours/distance.py
index f850648..8aa67a6 100644
--- a/neighbours/distance.py
+++ b/neighbours/distance.py
@@ -2,12 +2,22 @@ import numpy as np
 
 
 def euclidean(x: np.ndarray, y: np.ndarray):
+    """Calculate the Euclidean distance between two vectors
+
+    :param x: vector as np.ndarray
+    :param y: vector as np.ndarray
+    :return: Euclidean distance between two vectors
+    """
+
     return np.linalg.norm(x - y)
 
 
 def manhattan(x: np.ndarray, y: np.ndarray):
+    """Calculate the Manhattan distance between two vectors
+
+    :param x: vector as np.ndarray
+    :param y: vector as np.ndarray
+    :return: Manhattan distance between two vectors
+    """
+
     return np.sum(np.abs(x - y))
-
-
-def cosine(x: np.ndarray, y: np.ndarray):
-    return 1 - (x.dot(y) / (np.linalg.norm(x) * np.linalg.norm(y)))
diff --git a/neighbours/knn_classifier.py b/neighbours/knn_classifier.py
index 1a6048c..4a8ed71 100644
--- a/neighbours/knn_classifier.py
+++ b/neighbours/knn_classifier.py
@@ -5,13 +5,40 @@ from .exceptions import *
 
 
 class KNNClassifier:
+    """K-nearest neighbors classifier
+
+    Weighted kNN classifier based on random projection forest.
+
+    Supports different (including custom) smoothing kernels and distance metrics.
+
+    Attributes:
+        features: number of features in each sample
+        forest: an instance of RPTForest
+        classes: an array of labels (integers from 0 to N) corresponding to loaded train points
+        classes_count: number of classes used
+    """
+
     def __init__(self, features, classes_count, trees_count, rpt_m):
+        """Initializes new classifier
+
+        :param features: number of features in each sample
+        :param classes_count: number of classes used
+        :param trees_count: number of trees in the forest
+        :param rpt_m: maximum number of samples in one leaf of an RP tree
+        """
+
         self.features = features
         self.forest = RPTForest(features, trees_count, rpt_m)
         self.classes = None
         self.classes_count = classes_count
 
     def load(self, points, classes):
+        """Loads train data, builds a corresponding forest
+
+        :param points: np.ndarray of train samples
+        :param classes: an array of labels (integers from 0 to N) corresponding to loaded train points
+        """
+
         if not isinstance(points, np.ndarray):
             raise InvalidType("points should be represented as np.ndarray")
 
@@ -31,6 +58,15 @@ class KNNClassifier:
         self.forest.load(points)
 
     def predict(self, point: np.ndarray, distance, kernel, h):
+        """Predict class of given sample
+
+        :param point: target point as np.ndarray
+        :param distance: distance metric function (from neighbours.distance)
+        :param kernel: smoothing kernel function (from neighbours.kernel)
+        :param h: bandwidth
+        :return: predicted class
+        """
+
         nearest_point_indexes = self.forest.get_neighbours(point)
 
         votes = np.zeros(self.classes_count)