August 6, 2020

K-Nearest Neighbors Classifier

1. KNN is a classification algorithm. Data points with similar attributes fall into similar categories.

2. Three steps of the KNN algorithm:

   1. Normalize the data.

   2. Find the K-Nearest Neighbors.

   3. Classify the new point based on those neighbors.

3. Finding K-Nearest Neighbors.

   1. def distance(movie1, movie2):

      1. squared_difference = 0

      2. for i in range(len(movie1)):

         1. squared_difference += (movie1[i] - movie2[i]**2

      3. final_distance = squared_distance ** 0.5

      4. return final_distance

   2. def classify(unknown, dataset, k):

      1. distances = [ ]

      2. for title in dataset:

         1. movie = dataset[title]

         2. distance_to_point = distance(movie, unknown)

         3. #adding the distance and point associated

         4. distances.append([distance_to_point, title])

      3. distances.sort( ) #using Python’s sort function

      4. neighbors = distances[0:k] #only the k closest points

      5. return neighbors