引言

利用jupyter notebook实现一个简单的推荐系统，数据来源是Moivelen中的5M压缩文件

数据导入

准备两份数据来进行训练

（1）数据是拥有1682份数据的电影集

（2）用户点评数据

(3)电影明细表

#先把两份文件的数据读到文件内存里，用pandas来做数据读取。
import pandas as pd
import numpy as np
from sklearn.metrics.pairwise  import cosine_similarity
import time

def init():
    #初始化，读入文件：return：用户表，评分表，电影明细表
    user_file="E:\\Mycode\\Recommended_system\\ml-100k\\u.user"
    data_file="E:\\Mycode\\Recommended_system\\ml-100k\\u.data"
    item_file ="E:\\Mycode\\Recommended_system\\ml-100k\\u.item"
    u_cols = ['user_id', 'age', 'sex', 'occupation', 'zip_code']
    users = pd.read_csv(user_file, sep="|", names=u_cols, encoding='latin-1')
    r_cols = ['user_id', 'movie_id', 'rating', 'unix_timestamp']
    ratings = pd.read_csv(data_file, sep='\t', names=r_cols,encoding='latin-1')
    i_cols = ['movie_id', 'movie_title' ,'release date','video release date', 'IMDb URL', 'unknown',
              'Action', 'Adventure', 'Animation', 'Children\'s', 'Comedy', 'Crime', 'Documentary',
              'Drama', 'Fantasy', 'Film-Noir', 'Horror', 'Musical', 'Mystery', 'Romance', 'Sci-Fi',
              'Thriller', 'War', 'Western']
    items = pd.read_csv(item_file, sep='|', names=i_cols, encoding='latin-1')
    return users, ratings, items

users.head()

---------------------------------------------------------------------------

NameError                                 Traceback (most recent call last)

<ipython-input-2-505e5cd46540> in <module>
----> 1 users.head()


NameError: name 'users' is not defined

user_file="E:\\Mycode\\Recommended_system\\ml-100k\\u.user"
u_cols = ['user_id', 'age', 'sex', 'occupation', 'zip_code']
users = pd.read_csv(user_file, sep="|", names=u_cols, encoding='latin-1')

users

	user_id	age	sex	occupation	zip_code
0	1	24	M	technician	85711
1	2	53	F	other	94043
2	3	23	M	writer	32067
3	4	24	M	technician	43537
4	5	33	F	other	15213
5	6	42	M	executive	98101
6	7	57	M	administrator	91344
7	8	36	M	administrator	05201
8	9	29	M	student	01002
9	10	53	M	lawyer	90703
10	11	39	F	other	30329
11	12	28	F	other	06405
12	13	47	M	educator	29206
13	14	45	M	scientist	55106
14	15	49	F	educator	97301
15	16	21	M	entertainment	10309
16	17	30	M	programmer	06355
17	18	35	F	other	37212
18	19	40	M	librarian	02138
19	20	42	F	homemaker	95660
20	21	26	M	writer	30068
21	22	25	M	writer	40206
22	23	30	F	artist	48197
23	24	21	F	artist	94533
24	25	39	M	engineer	55107
25	26	49	M	engineer	21044
26	27	40	F	librarian	30030
27	28	32	M	writer	55369
28	29	41	M	programmer	94043
29	30	7	M	student	55436
...	...	...	...	...	...
913	914	44	F	other	08105
914	915	50	M	entertainment	60614
915	916	27	M	engineer	N2L5N
916	917	22	F	student	20006
917	918	40	M	scientist	70116
918	919	25	M	other	14216
919	920	30	F	artist	90008
920	921	20	F	student	98801
921	922	29	F	administrator	21114
922	923	21	M	student	E2E3R
923	924	29	M	other	11753
924	925	18	F	salesman	49036
925	926	49	M	entertainment	01701
926	927	23	M	programmer	55428
927	928	21	M	student	55408
928	929	44	M	scientist	53711
929	930	28	F	scientist	07310
930	931	60	M	educator	33556
931	932	58	M	educator	06437
932	933	28	M	student	48105
933	934	61	M	engineer	22902
934	935	42	M	doctor	66221
935	936	24	M	other	32789
936	937	48	M	educator	98072
937	938	38	F	technician	55038
938	939	26	F	student	33319
939	940	32	M	administrator	02215
940	941	20	M	student	97229
941	942	48	F	librarian	78209
942	943	22	M	student	77841

943 rows × 5 columns

data_file="E:\\Mycode\\Recommended_system\\ml-100k\\u.data"
r_cols = ['user_id', 'movie_id', 'rating', 'unix_timestamp']
ratings = pd.read_csv(data_file, sep='\t', names=r_cols,encoding='latin-1')
ratings

	user_id	movie_id	rating	unix_timestamp
0	196	242	3	881250949
1	186	302	3	891717742
2	22	377	1	878887116
3	244	51	2	880606923
4	166	346	1	886397596
5	298	474	4	884182806
6	115	265	2	881171488
7	253	465	5	891628467
8	305	451	3	886324817
9	6	86	3	883603013
10	62	257	2	879372434
11	286	1014	5	879781125
12	200	222	5	876042340
13	210	40	3	891035994
14	224	29	3	888104457
15	303	785	3	879485318
16	122	387	5	879270459
17	194	274	2	879539794
18	291	1042	4	874834944
19	234	1184	2	892079237
20	119	392	4	886176814
21	167	486	4	892738452
22	299	144	4	877881320
23	291	118	2	874833878
24	308	1	4	887736532
25	95	546	2	879196566
26	38	95	5	892430094
27	102	768	2	883748450
28	63	277	4	875747401
29	160	234	5	876861185
...	...	...	...	...
99970	449	120	1	879959573
99971	661	762	2	876037121
99972	721	874	3	877137447
99973	821	151	4	874792889
99974	764	596	3	876243046
99975	537	443	3	886031752
99976	618	628	2	891308019
99977	487	291	3	883445079
99978	113	975	5	875936424
99979	943	391	2	888640291
99980	864	685	4	888891900
99981	750	323	3	879445877
99982	279	64	1	875308510
99983	646	750	3	888528902
99984	654	370	2	887863914
99985	617	582	4	883789294
99986	913	690	3	880824288
99987	660	229	2	891406212
99988	421	498	4	892241344
99989	495	1091	4	888637503
99990	806	421	4	882388897
99991	676	538	4	892685437
99992	721	262	3	877137285
99993	913	209	2	881367150
99994	378	78	3	880056976
99995	880	476	3	880175444
99996	716	204	5	879795543
99997	276	1090	1	874795795
99998	13	225	2	882399156
99999	12	203	3	879959583

100000 rows × 4 columns

item_file ="E:\\Mycode\\Recommended_system\\ml-100k\\u.item"
i_cols = ['movie_id', 'movie_title' ,'release date','video release date', 'IMDb URL', 'unknown',
              'Action', 'Adventure', 'Animation', 'Children\'s', 'Comedy', 'Crime', 'Documentary',
              'Drama', 'Fantasy', 'Film-Noir', 'Horror', 'Musical', 'Mystery', 'Romance', 'Sci-Fi',
              'Thriller', 'War', 'Western']
items = pd.read_csv(item_file, sep='|', names=i_cols, encoding='latin-1')
items

	movie_id	movie_title	release date	video release date	IMDb URL	unknown	Action	Adventure	Animation	Children's	...	Fantasy	Film-Noir	Horror	Musical	Mystery	Romance	Sci-Fi	Thriller	War	Western
0	1	Toy Story (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Toy%20Story%2...	0	0	0	1	1	...	0	0	0	0	0	0	0	0	0	0
1	2	GoldenEye (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?GoldenEye%20(...	0	1	1	0	0	...	0	0	0	0	0	0	0	1	0	0
2	3	Four Rooms (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Four%20Rooms%...	0	0	0	0	0	...	0	0	0	0	0	0	0	1	0	0
3	4	Get Shorty (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Get%20Shorty%...	0	1	0	0	0	...	0	0	0	0	0	0	0	0	0	0
4	5	Copycat (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Copycat%20(1995)	0	0	0	0	0	...	0	0	0	0	0	0	0	1	0	0
5	6	Shanghai Triad (Yao a yao yao dao waipo qiao) ...	01-Jan-1995	NaN	http://us.imdb.com/Title?Yao+a+yao+yao+dao+wai...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
6	7	Twelve Monkeys (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Twelve%20Monk...	0	0	0	0	0	...	0	0	0	0	0	0	1	0	0	0
7	8	Babe (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Babe%20(1995)	0	0	0	0	1	...	0	0	0	0	0	0	0	0	0	0
8	9	Dead Man Walking (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Dead%20Man%20...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
9	10	Richard III (1995)	22-Jan-1996	NaN	http://us.imdb.com/M/title-exact?Richard%20III...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	1	0
10	11	Seven (Se7en) (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Se7en%20(1995)	0	0	0	0	0	...	0	0	0	0	0	0	0	1	0	0
11	12	Usual Suspects, The (1995)	14-Aug-1995	NaN	http://us.imdb.com/M/title-exact?Usual%20Suspe...	0	0	0	0	0	...	0	0	0	0	0	0	0	1	0	0
12	13	Mighty Aphrodite (1995)	30-Oct-1995	NaN	http://us.imdb.com/M/title-exact?Mighty%20Aphr...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
13	14	Postino, Il (1994)	01-Jan-1994	NaN	http://us.imdb.com/M/title-exact?Postino,%20Il...	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
14	15	Mr. Holland's Opus (1995)	29-Jan-1996	NaN	http://us.imdb.com/M/title-exact?Mr.%20Holland...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
15	16	French Twist (Gazon maudit) (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Gazon%20maudi...	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
16	17	From Dusk Till Dawn (1996)	05-Feb-1996	NaN	http://us.imdb.com/M/title-exact?From%20Dusk%2...	0	1	0	0	0	...	0	0	1	0	0	0	0	1	0	0
17	18	White Balloon, The (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Badkonake%20S...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
18	19	Antonia's Line (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Antonia%20(1995)	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
19	20	Angels and Insects (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Angels%20and%...	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
20	21	Muppet Treasure Island (1996)	16-Feb-1996	NaN	http://us.imdb.com/M/title-exact?Muppet%20Trea...	0	1	1	0	0	...	0	0	0	1	0	0	0	1	0	0
21	22	Braveheart (1995)	16-Feb-1996	NaN	http://us.imdb.com/M/title-exact?Braveheart%20...	0	1	0	0	0	...	0	0	0	0	0	0	0	0	1	0
22	23	Taxi Driver (1976)	16-Feb-1996	NaN	http://us.imdb.com/M/title-exact?Taxi%20Driver...	0	0	0	0	0	...	0	0	0	0	0	0	0	1	0	0
23	24	Rumble in the Bronx (1995)	23-Feb-1996	NaN	http://us.imdb.com/M/title-exact?Hong%20Faan%2...	0	1	1	0	0	...	0	0	0	0	0	0	0	0	0	0
24	25	Birdcage, The (1996)	08-Mar-1996	NaN	http://us.imdb.com/M/title-exact?Birdcage,%20T...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
25	26	Brothers McMullen, The (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Brothers%20Mc...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
26	27	Bad Boys (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Bad%20Boys%20...	0	1	0	0	0	...	0	0	0	0	0	0	0	0	0	0
27	28	Apollo 13 (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Apollo%2013%2...	0	1	0	0	0	...	0	0	0	0	0	0	0	1	0	0
28	29	Batman Forever (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Batman%20Fore...	0	1	1	0	0	...	0	0	0	0	0	0	0	0	0	0
29	30	Belle de jour (1967)	01-Jan-1967	NaN	http://us.imdb.com/M/title-exact?Belle%20de%20...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1652	1653	Entertaining Angels: The Dorothy Day Story (1996)	27-Sep-1996	NaN	http://us.imdb.com/M/title-exact?Entertaining%...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1653	1654	Chairman of the Board (1998)	01-Jan-1998	NaN	http://us.imdb.com/Title?Chairman+of+the+Board...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1654	1655	Favor, The (1994)	01-Jan-1994	NaN	http://us.imdb.com/M/title-exact?Favor,%20The%...	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
1655	1656	Little City (1998)	20-Feb-1998	NaN	http://us.imdb.com/M/title-exact?Little+City+(...	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
1656	1657	Target (1995)	28-Feb-1996	NaN	http://us.imdb.com/M/title-exact?Target%20(1995)	0	1	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1657	1658	Substance of Fire, The (1996)	06-Dec-1996	NaN	http://us.imdb.com/M/title-exact?Substance%20o...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1658	1659	Getting Away With Murder (1996)	12-Apr-1996	NaN	http://us.imdb.com/Title?Getting+Away+With+Mur...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1659	1660	Small Faces (1995)	09-Aug-1996	NaN	http://us.imdb.com/M/title-exact?Small%20Faces...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1660	1661	New Age, The (1994)	01-Jan-1994	NaN	http://us.imdb.com/M/title-exact?New%20Age,%20...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1661	1662	Rough Magic (1995)	30-May-1997	NaN	http://us.imdb.com/M/title-exact?Rough%20Magic...	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
1662	1663	Nothing Personal (1995)	30-Apr-1997	NaN	http://us.imdb.com/M/title-exact?Nothing%20Per...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	1	0
1663	1664	8 Heads in a Duffel Bag (1997)	18-Apr-1997	NaN	http://us.imdb.com/Title?8+Heads+in+a+Duffel+B...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1664	1665	Brother's Kiss, A (1997)	25-Apr-1997	NaN	http://us.imdb.com/M/title-exact?Brother%27s%2...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1665	1666	Ripe (1996)	02-May-1997	NaN	http://us.imdb.com/M/title-exact?Ripe%20%28199...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1666	1667	Next Step, The (1995)	13-Jun-1997	NaN	http://us.imdb.com/M/title-exact?Next%20Step%2...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1667	1668	Wedding Bell Blues (1996)	13-Jun-1997	NaN	http://us.imdb.com/M/title-exact?Wedding%20Bel...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1668	1669	MURDER and murder (1996)	20-Jun-1997	NaN	http://us.imdb.com/M/title-exact?MURDER+and+mu...	0	0	0	0	0	...	0	0	0	0	1	0	0	0	0	0
1669	1670	Tainted (1998)	01-Feb-1998	NaN	http://us.imdb.com/M/title-exact?Tainted+(1998)	0	0	0	0	0	...	0	0	0	0	0	0	0	1	0	0
1670	1671	Further Gesture, A (1996)	20-Feb-1998	NaN	http://us.imdb.com/M/title-exact?Further+Gestu...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1671	1672	Kika (1993)	01-Jan-1993	NaN	http://us.imdb.com/M/title-exact?Kika%20(1993)	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1672	1673	Mirage (1995)	01-Jan-1995	NaN	http://us.imdb.com/M/title-exact?Mirage%20(1995)	0	1	0	0	0	...	0	0	0	0	0	0	0	1	0	0
1673	1674	Mamma Roma (1962)	01-Jan-1962	NaN	http://us.imdb.com/M/title-exact?Mamma%20Roma%...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1674	1675	Sunchaser, The (1996)	25-Oct-1996	NaN	http://us.imdb.com/M/title-exact?Sunchaser,%20...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1675	1676	War at Home, The (1996)	01-Jan-1996	NaN	http://us.imdb.com/M/title-exact?War%20at%20Ho...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1676	1677	Sweet Nothing (1995)	20-Sep-1996	NaN	http://us.imdb.com/M/title-exact?Sweet%20Nothi...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1677	1678	Mat' i syn (1997)	06-Feb-1998	NaN	http://us.imdb.com/M/title-exact?Mat%27+i+syn+...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1678	1679	B. Monkey (1998)	06-Feb-1998	NaN	http://us.imdb.com/M/title-exact?B%2E+Monkey+(...	0	0	0	0	0	...	0	0	0	0	0	1	0	1	0	0
1679	1680	Sliding Doors (1998)	01-Jan-1998	NaN	http://us.imdb.com/Title?Sliding+Doors+(1998)	0	0	0	0	0	...	0	0	0	0	0	1	0	0	0	0
1680	1681	You So Crazy (1994)	01-Jan-1994	NaN	http://us.imdb.com/M/title-exact?You%20So%20Cr...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1681	1682	Scream of Stone (Schrei aus Stein) (1991)	08-Mar-1996	NaN	http://us.imdb.com/M/title-exact?Schrei%20aus%...	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0

1682 rows × 24 columns

构建用户矩阵

# 用户总数
num_users = users.user_id.unique().shape[0]
# 电影总数
num_items = ratings.movie_id.unique().shape[0]
data_matrix=np.zeros((num_users,num_items))

def constructUserMovieMatrix(users, ratings):
    '''
    构造用户-电影矩阵
    :param users: 用户表
    :param ratings: 打分表
    :return: 用户对电影评分的矩阵
    '''
    num_users = users.user_id.unique().shape[0]   #用户数
    num_items = ratings.movie_id.unique().shape[0]   #电影总数
    data_matrix = np.zeros((num_users, num_items))
    for line in ratings.itertuples():
        data_matrix[line[1]-1, line[2]-1] = line[3]
    return data_matrix

constructUserMovieMatrix(users, ratings)

array([[5., 3., 4., ..., 0., 0., 0.],
       [4., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       ...,
       [5., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 5., 0., ..., 0., 0., 0.]])

1.unique()方法去重；
2.shape[0]返回表示DataFrame维度的元祖；
3.line的内容

def calculationSimilarity(data_matrix):
    '''
    转置计算电影之间相似度矩阵，不转置计算用户之间相似度矩阵
    :param data_matrix: 评分矩阵
    :return: 电影之间的相似度矩阵
    '''
    user_similarity = cosine_similarity(data_matrix, dense_output=True)
    item_similarity = cosine_similarity(data_matrix.T, dense_output=True)
    return item_similarity

def rec_sys(items,ratings, item_similarity, keywords, k):
    '''
    推荐系统
    :param items: 电影明细表
    :param ratings: 评分表
    :param item_similarity: 电影相似度矩阵
    :param keywords: 输入的电影名称或关键字
    :param k: 推荐个数
    :return: 推荐电影结果列表
    '''
    movie_list = []     # 存储推荐电影结果列表
    movie_id = list(items[items['movie_title'].str.contains(keywords)].movie_id)[0]   # 获得电影的id
    movie_similarity = item_similarity[movie_id - 1]    # 计算该电影的余弦相似度数组
    movie_similarity_index = np.argsort(-movie_similarity)[1:k + 1]     # 返回前k+1个最高相似度的索引位置
    for index in movie_similarity_index:
        rec_movie = list(items[items['movie_id'] == index + 1].movie_title)     # 电影名
        rec_movie.append(movie_similarity[index])    # 相似度
        rec_movie.append(ratings[ratings['movie_id'] == index+1].rating.mean()) # 平均评分
        rec_movie.append(len(ratings[ratings['movie_id'] == index+1]))    # 评分用户数
        movie_list.append(rec_movie)
    return movie_list

if __name__ == '__main__':
    beginTime = time.time()
    keywords = "Assassins"
    k = 5
    keywords = keywords.title()
    users, ratings, items = init()
    data_matrix = constructUserMovieMatrix(users, ratings)
    similarity = calculationSimilarity(data_matrix)
    movie_list = rec_sys(items, ratings, similarity, keywords, k)
    print(movie_list)
    print("推荐耗时：", time.time()-beginTime)

[['Outbreak (1995)', 0.477148560717635, 3.2403846153846154, 104], ['Client, The (1994)', 0.4185073174104336, 3.381443298969072, 97], ['Net, The (1995)', 0.40070132781155243, 3.0083333333333333, 120], ['Under Siege 2: Dark Territory (1995)', 0.3980230845871082, 2.4583333333333335, 48], ['Demolition Man (1993)', 0.39672213644098525, 3.152173913043478, 92]]
推荐耗时： 0.5808792114257812

参考链接

博客

一个简单的推荐系统实现

引言

数据导入

构建用户矩阵

参考链接

猜你喜欢