基于协同过滤的物联网场景个性化推荐系统

基于MovieLens数据集的协同过滤推荐系统

from tkinter import *
import tkinter.font as tkFont
import threading
import os
from operator import itemgetter, attrgetter  
from math import sqrt
import time


def load_data():  
    filename_user_movie = 'data/u.data'  
    filename_movieInfo = 'data/u.item'  
    user_movie = {}  
    for line in open(filename_user_movie,encoding='ISO-8859-1'):  
        (userId, itemId, rating, timestamp) = line.strip().split('\t')  
        user_movie.setdefault(userId,{})  
        user_movie[userId][itemId] = float(rating)  
      
    movies = {}  
    for line in open(filename_movieInfo,encoding='ISO-8859-1'):  
        (movieId, movieTitle) = line.split('|')[0:2]  
        movies[movieId] = movieTitle  
      
    return user_movie, movies  

def average_rating(user):  
    average = 0  
    for u in user_movie[user].keys():  
        average += user_movie[user][u]  
    average = average * 1.0 / len(user_movie[user].keys())  
    return average  
  
def calUserSim(user_movie):  
    # build inverse table for movie_user  
    movie_user = {}  
    for ukey in user_movie.keys():  
        for mkey in user_movie[ukey].keys():  
            if mkey not in movie_user:  
                movie_user[mkey] = []  
            movie_user[mkey].append(ukey)  
  
    # calculated co-rated movies between users  
    C = {}  
    for movie, users in movie_user.items():  
        for u in users:  
            C.setdefault(u,{})  
            for n in users:  
                if u == n:  
                    continue  
                C[u].setdefault(n,[])  
                C[u][n].append(movie)  
                  
    # calculate user similarity (perason correlation)  
    userSim = {}  
    for u in C.keys():  
          
        for n in C[u].keys():  
              
            userSim.setdefault(u,{})  
            userSim[u].setdefault(n,0)  
          
            average_u_rate = average_rating(u)  
            average_n_rate = average_rating(n)  
              
            part1 = 0  
            part2 = 0  
            part3 = 0  
            for m in C[u][n]:  
                part1 += (user_movie[u][m]-average_u_rate)*(user_movie[n][m]-average_n_rate)*1.0  
                part2 += pow(user_movie[u][m]-average_u_rate, 2)*1.0  
                part3 += pow(user_movie[n][m]-average_n_rate, 2)*1.0              
            part2 = sqrt(part2)  
            part3 = sqrt(part3)  
            if part2 == 0:  
                part2 = 0.001  
            if part3 == 0:  
                part3 = 0.001   
            userSim[u][n] = part1 / (part2 * part3)     
    return userSim  
  
def getRecommendations(user, user_movie, movies, userSim, N):  
    pred = {}  
    interacted_items = user_movie[user].keys()  
    average_u_rate = average_rating(user)  
    sumUserSim = 0  
    for n, nuw in sorted(userSim[user].items(),key=itemgetter(1),reverse=True)[0:N]:  
        average_n_rate = average_rating(n)  
        for i, nrating in user_movie[n].items():  
            # filter movies user interacted before  
            if i in interacted_items:  
                continue  
            pred.setdefault(i,0)  
            pred[i] += nuw * (nrating - average_n_rate)  
        sumUserSim += nuw  
  
    for i, rating in pred.items():  
        pred[i] = average_u_rate + (pred[i]*1.0) / sumUserSim  
          
    # top-10 pred  
    pred = sorted(pred.items(), key=itemgetter(1), reverse=True)[0:10]  
    return pred 

def thread_it(func,*args):
    """将函数打包进线程内执行"""
    # 创建
    t = threading.Thread(target=func, args=args) 
    # 守护 
    t.setDaemon(True) 
    # 启动
    t.start()

user_movie, movies = load_data()  
  

def star():
    text2.delete(1.0,END)
    text2.insert(1.0,'正在为该用户推荐中，请稍等...')
    t=0
    while t<40:
        time.sleep(0.5)
        text2.insert(3.0,'.')
        t+=1
    # Calculate user similarity  
    userSim = calUserSim(user_movie)
    # Recommend
    n=text1.get(1.0,END)
    n=str.strip(n) 
    pred = getRecommendations('%s'%n, user_movie, movies, userSim, 20)  
    # display recommend result (top-10 results)
    text2.delete(1.0,END)
    text2.insert(1.0,'推荐结果如下：\n\n')
    for i, rating in pred:        
        #print('电影名: %s,  评分: %s' % (movies[i], rating) )
        text2.insert(1.0,'电影名: %s,  评分: %s' % (movies[i], rating)+'\n\n')


#实例化一个父界面
init_window = Tk()
#界面参数
init_window.title('电影推荐系统')           #窗口名称
init_window.geometry('600x500+400+150')    #窗口大小
init_window["bg"] = "skyblue"              #窗口颜色
init_window.attributes("-alpha",0.8)       #窗口透明度
init_window.iconbitmap("1.ico")            #窗口图标
#标签
label1=Label(init_window,text="请输入用户ID",bg='skyblue',font=('隶书',15))
label1.place(relx=0.36,rely=0.05)
label2=Label(init_window,text="推荐结果",bg='skyblue',font=('隶书',15))
label2.place(relx=0.4,rely=0.37)
#文本框
text1=Text(init_window,width=60, height=4)
text1.place(relx=0.15,rely=0.12)
text2=Text(init_window,width=70, height=15,fg='steelblue')
text2.place(relx=0.1,rely=0.42)
#按钮
button1=Button(init_window,text="确定", bg="Cornsilk", width=8,height=1,command=lambda :thread_it(star))
button1.place(relx=0.44,rely=0.25)

#窗口进入循环
init_window.mainloop()