推荐引擎

推荐引擎

欧氏距离分数

通过欧氏距离判断相关性，判断相似程度
$$欧氏距离分数=\frac{1}{1+欧氏距离}$$
欧氏距离->0 - oo
欧氏距离分数->1 - 0

     0 <-> 1
不相似   相似
    a   b   c   ...
a  x   x   x   ...
b  x   x   x   ...
c  x   x   x   ...
...

代码：

import json
import numpy as np
with open(r'C:\Users\Cs\Desktop\机器学习\ML\data\ratings.json', 'r') as f:
    ratings = json.loads(f.read())
users, scmat = list(ratings.keys()), []
for user1 in users:
    scrow = []
    for user2 in users:
        movies = set()
        for movie in ratings[user1]:
            if movie in ratings[user2]:
                movies.add(movie)
        if len(movies) == 0:
            score = 0
        else:
            x, y = [], []
            for movie in movies:
                x.append(ratings[user1][movie])
                y.append(ratings[user2][movie])
            x = np.array(x)
            y = np.array(y)
            score = 1 / (1 + np.sqrt(
                ((x - y) ** 2).sum()))
        scrow.append(score)
    scmat.append(scrow)
users = np.array(users)
scmat = np.array(scmat)
for scrow in scmat:
    print(' '.join('{:>5.2f}'.format(score)
                   for score in scrow))

皮氏距离分数

通过皮氏距离相关性判断两者喜欢的电影的程度。
相关性矩阵
/ 1 相关性系数
\ 相关性系数 1 /
相关性系数 = 协方差/标准差之积
-1 <------- 0 -------> 1
反相关 不相关 正相关
3-5
0-2
代码：

import json
import numpy as np
with open(r'C:\Users\Cs\Desktop\机器学习\ML\data\ratings.json', 'r') as f:
    ratings = json.loads(f.read())
users, scmat = list(ratings.keys()), []
for user1 in users:
    scrow = []
    for user2 in users:
        movies = set()
        for movie in ratings[user1]:
            if movie in ratings[user2]:
                movies.add(movie)
        if len(movies) == 0:
            score = 0
        else:
            x, y = [], []
            for movie in movies:
                x.append(ratings[user1][movie])
                y.append(ratings[user2][movie])
            x = np.array(x)
            y = np.array(y)
            score = np.corrcoef(x, y)[0, 1]
        scrow.append(score)
    scmat.append(scrow)
users = np.array(users)
scmat = np.array(scmat)
for scrow in scmat:
    print(' '.join('{:>5.2f}'.format(score)
                   for score in scrow))

按照相似度从高到低排列每个用户的相似用户

代码：

# -*- coding: utf-8 -*-
from __future__ import unicode_literals
import json
import numpy as np
with open(r'C:\Users\Cs\Desktop\机器学习\ML\data\ratings.json', 'r') as f:
    ratings = json.loads(f.read())
users, scmat = list(ratings.keys()), []
#遍历两次users列表，匹配每个人和所有人的皮氏距离分数。并记录每个人的分数列表和下标
for user1 in users:
	# 记录每个人的分数列表，
    scrow = []
    for user2 in users:
    	# 记录两个人共同的电影集合
        movies = set()
        for movie in ratings[user1]:
            if movie in ratings[user2]:
                movies.add(movie)
        # 这里其实有一个问题，没有考虑两个人的电影集合问题，建议分数根据相似电影占总电影比例进行一次加权。
        if len(movies) == 0:
            score = 0
        else:
            x, y = [], []
            for movie in movies:
                x.append(ratings[user1][movie])
                y.append(ratings[user2][movie])
            x = np.array(x)
            y = np.array(y)
            score = np.corrcoef(x, y)[0, 1]
        scrow.append(score)
    scmat.append(scrow)
users = np.array(users)
scmat = np.array(scmat)
# 按照下标顺序得到每个人的分数并排序输出
for i, user in enumerate(users):
    sorted_indices = scmat[i].argsort()[::-1]
    sorted_indices = sorted_indices[
        sorted_indices != i]
    similar_users = users[sorted_indices]
    similar_scores = scmat[i, sorted_indices]
    print(user, similar_users, similar_scores,
          sep='\n')

4.生成推荐清单
推荐度
皮氏距离分数>0的用户
打分高低
相似度权重
代码：

# -*- coding: utf-8 -*-
from __future__ import unicode_literals
import json
import numpy as np
with open('../../data/ratings.json', 'r') as f:
    ratings = json.loads(f.read())
users, scmat = list(ratings.keys()), []
for user1 in users:
    scrow = []
    for user2 in users:
        movies = set()
        for movie in ratings[user1]:
            if movie in ratings[user2]:
                movies.add(movie)
        if len(movies) == 0:
            score = 0
        else:
            x, y = [], []
            for movie in movies:
                x.append(ratings[user1][movie])
                y.append(ratings[user2][movie])
            x = np.array(x)
            y = np.array(y)
            score = np.corrcoef(x, y)[0, 1]
        scrow.append(score)
    scmat.append(scrow)
users = np.array(users)
scmat = np.array(scmat)
for i, user in enumerate(users):
    sorted_indices = scmat[i].argsort()[::-1]
    sorted_indices = sorted_indices[
        sorted_indices != i]
    similar_users = users[sorted_indices]
    similar_scores = scmat[i, sorted_indices]
    positive_mask = similar_scores > 0
    similar_users = similar_users[positive_mask]
    similar_scores = similar_scores[positive_mask]
    score_sums, weight_sums = {}, {}
    for similar_user, similar_score in zip(
            similar_users, similar_scores):
        for movie, score in ratings[
                similar_user].items():
            if movie not in ratings[user].keys():
                if movie not in score_sums.keys():
                    score_sums[movie] = 0
                score_sums[movie] += score * similar_score
                if movie not in weight_sums.keys():
                    weight_sums[movie] = 0
                weight_sums[movie] += similar_score
    movie_ranks = {}
    for movie, score_sum in score_sums.items():
        movie_ranks[movie] = \
            score_sum / weight_sums[movie]
    sorted_indices = np.array(list(
        movie_ranks.values())).argsort()[::-1]
    recomms = np.array(list(
        movie_ranks.keys()))[sorted_indices]
    print(user, recomms, sep='\n')