knn 畢式定理在高維空間

原始碼

import math

film = [5, 7, 8, 10, 2]             # 玩命關頭特徵值
film_titles = [                     # 比較影片片名
    '復仇者聯盟',
    '決戰中途島',
    '冰雪奇緣',
    '雙子殺手',
]
film_features = [                   # 比較影片特徵值
    [2, 8, 8, 5, 6],
    [5, 6, 9, 2, 5],
    [8, 2, 0, 0, 10],
    [5, 8, 8, 8, 3],
]

dist = []                           # 儲存影片相似度值
for f in film_features:
    distances = 0
    for i in range(len(f)):
        distances += (film[i] - f[i]) ** 2
    dist.append(math.sqrt(distances))
    
min = min(dist)                     # 求最小值
min_index = dist.index(min)         # 最小值的索引

print("與玩命關頭最相似的電影 : ", film_titles[min_index])
print("相似度值 : ", dist[min_index])
for i in range(len(dist)):
    print("影片 : %s, 相似度 : %6.2f" % (film_titles[i], dist[i]))

基本測試

import math

film = [5, 7, 8, 10, 2]             # 玩命關頭特徵值
film_titles = [                     # 比較影片片名
    '復仇者聯盟',
    '決戰中途島',
    '冰雪奇緣',
    '雙子殺手',
]
film_features = [                   # 比較影片特徵值
    [2, 8, 8, 5, 6],
    [5, 6, 9, 2, 5],
    [8, 2, 0, 0, 10],
    [5, 8, 8, 8, 3],
]
dist=[]
num=0
for f in film_features:
    print("這是file_features[{}]{}".format(num,f))
    num=num+1

計算list[0]的平方和

import math

film = [5, 7, 8, 10, 2]             # 玩命關頭特徵值
film_titles = [                     # 比較影片片名
    '復仇者聯盟',
    '決戰中途島',
    '冰雪奇緣',
    '雙子殺手',
]
film_features = [                   # 比較影片特徵值
    [2, 8, 8, 5, 6],
    [5, 6, 9, 2, 5],
    [8, 2, 0, 0, 10],
    [5, 8, 8, 8, 3],
]
distances=0
for i in range(len(film)):
    distances=distances+(film[i]-film_features[0][i])**2
a_squr=math.sqrt(distances)
print(a_squr)

使用二層迴圈計算

import math

film = [5, 7, 8, 10, 2]             # 玩命關頭特徵值
film_titles = [                     # 比較影片片名
    '復仇者聯盟',
    '決戰中途島',
    '冰雪奇緣',
    '雙子殺手',
]
film_features = [                   # 比較影片特徵值
    [2, 8, 8, 5, 6],
    [5, 6, 9, 2, 5],
    [8, 2, 0, 0, 10],
    [5, 8, 8, 8, 3],
]
dist = []                           # 儲存影片相似度值
for f in film_features:
    distances = 0
    for i in range(len(f)):
        distances += (film[i] - f[i]) ** 2
    dist.append(math.sqrt(distances))
    print(dist)
print(type(dist))
print(type(distances))