def predict(self, x):
"""Predict x array
:param x: data with coordinates to be predicted
:type: list, tuple, np.array
:example:
[[1, 2], [2, 3], [3, 4], [4, 5]]
:return: x predicts
:type: list
:example:
[[1], [0], [0], [1]]
"""
predicts = []
for coordinates in x:
predict_point = Point(coordinates)
# euclidean distance from predict_point to all in self._fit_data
distances = []
for data_point, data_label in self._fit_data:
distances.append((predict_point.distance(data_point), data_label))
# k points with less distances
distances = sorted(distances, key=itemgetter(0))[:self.k]
# label of k points with less distances
predicts.append(list(max(distances, key=itemgetter(1))[1]))
return predicts
def pairwise_distances(points, centroids):
"""Create pairwise distance between points and the centroids
:param points:
:type: list, tuple, np.array
:example:
[[1, 2], [2, 3], [3, 4], [4, 5]]
:param centroids:
:type: list, tuple, np.array
:example:
[[0, 0], [3, 3]]
:return: distance
:type: list
:example:
[[14.866068747318506, 1.0, 0.0],
[14.212670403551895, 1.4142135623730951, 1.0],
[14.142135623730951, 0.0, 1.0], ... ]
"""
distances = list()
for index, point in enumerate(points):
distances.append([])
point = Point(point)
for centroid in centroids:
centroid = Point(centroid)
distance = centroid.distance(point)
distances[index].append(distance)
return distances