基于TensorFlow+CNN+协同过滤算法的智能电影推荐系统-深度学习算法应用(含微信小程序、ipynb工程源码)+数据集

import pandas as pd import numpy as np import pickle import random import os def get_random_movies(PROJECT_ROOT): title_count, title_set, genres2int, features, targets_values, ratings, users, movies, data, movies_orig, users_orig = pickle.load(open(os.path.join(PROJECT_ROOT,'preprocess.p'), mode='rb')) random_movies = [random.randint(0,3833) for i in range(5)] return movies_orig[random_movies] def get_a_movie(PROJECT_ROOT,movie_id): title_count, title_set, genres2int, features, targets_values, ratings, users, movies, data, movies_orig, users_orig = pickle.load(open(os.path.join(PROJECT_ROOT,'preprocess.p'), mode='rb')) #电影ID转下标的字典，数据集中电影ID跟下标不一致，比如第5行的数据电影ID不一定是5 movieid2idx = {val[0]:i for i, val in enumerate(movies.values)} this_movie_data = movies_orig[movieid2idx[movie_id]] return this_movie_data def recommend_same_type_movie(PROJECT_ROOT,movie_id_val, top_k=20): title_count, title_set, genres2int, features, targets_values, ratings, users, movies, data, movies_orig, users_orig = pickle.load(open(os.path.join(PROJECT_ROOT,'preprocess.p'), mode='rb')) #电影ID转下标的字典，数据集中电影ID跟下标不一致，比如第5行的数据电影ID不一定是5 movieid2idx = {val[0]:i for i, val in enumerate(movies.values)} movie_matrics = pickle.load(open(os.path.join(PROJECT_ROOT,'movie_matrics.p'), mode='rb')) users_matrics = pickle.load(open(os.path.join(PROJECT_ROOT,'users_matrics.p'), mode='rb')) # 推荐与选择类型同类型的电影 # print("您选择的电影是：{}".format(movies_orig[movieid2idx[movie_id_val]])) # 规范化电影特征矩阵 norm_movie_matrics = np.sqrt(np.sum(np.square(movie_matrics),axis=1)).reshape(3883,1) normalized_movie_matrics = movie_matrics/norm_movie_matrics # 获取所选电影特征向量 probs_embeddings = (movie_matrics[movieid2idx[movie_id_val]]).reshape([1, 200]) probs_embeddings = probs_embeddings/np.sqrt(np.sum(np.square(probs_embeddings))) # 计算相似度 probs_similarity = np.matmul(probs_embeddings, np.transpose(normalized_movie_matrics)) # print("根据您看的电影类型给您的推荐：") p = np.squeeze(probs_similarity) p[np.argsort(p)[:-top_k]] = 0 p = p / np.sum(p) results = set() while len(results) != 5: c = np.random.choice(3883, 1, p=p)[0] results.add(c) final_results = [movies_orig[val] for val in results] return final_results def recommend_other_favorite_movie(PROJECT_ROOT,movie_id_val, top_k = 20): title_count, title_set, genres2int, features, targets_values, ratings, users, movies, data, movies_orig, users_orig = pickle.load(open(os.path.join(PROJECT_ROOT,'preprocess.p'), mode='rb')) #电影ID转下标的字典，数据集中电影ID跟下标不一致，比如第5行的数据电影ID不一定是5 movieid2idx = {val[0]:i for i, val in enumerate(movies.values)} movie_matrics = pickle.load(open(os.path.join(PROJECT_ROOT,'movie_matrics.p'), mode='rb')) users_matrics = pickle.load(open(os.path.join(PROJECT_ROOT,'users_matrics.p'), mode='rb')) # 推荐看过同一个的电影的人喜欢的电影 # print("您看的电影是：{}".format(movies_orig[movieid2idx[movie_id_val]])) # 根据电影寻找相似的人 probs_movie_embeddings = (movie_matrics[movieid2idx[movie_id_val]]).reshape([1, 200]) probs_movie_embeddings = probs_movie_embeddings/np.sqrt(np.sum(np.square(probs_movie_embeddings))) norm_users_matrics = np.sqrt(np.sum(np.square(users_matrics),axis=1)).reshape(6040,1) normalized_users_matrics = users_matrics/norm_users_matrics probs_user_favorite_similarity = np.matmul(probs_movie_embeddings, np.transpose(normalized_users_matrics)) favorite_user_id = np.argsort(probs_user_favorite_similarity)[0][-top_k:] # print("喜欢看这个电影的人是：{}".format(users_orig[favorite_user_id-1])) # 他们喜欢什么样的电影 probs_users_embeddings = (users_matrics[favorite_user_id-1]).reshape([-1, 200]) probs_users_embeddings = probs_users_embeddings/np.sqrt(np.sum(np.square(probs_users_embeddings))) norm_movie_matrics = np.sqrt(np.sum(np.square(movie_matrics),axis=1)).reshape(3883,1) normalized_movie_matrics = movie_matrics/norm_movie_matrics probs_similarity = np.matmul(probs_users_embeddings, np.transpose(normalized_movie_matrics)) p = np.argmax(probs_similarity, 1) # print("喜欢看这个电影的人还喜欢看：") results = set() while len(results) != 5: c = p[random.randrange(top_k)] results.add(c) final_results = [movies_orig[val] for val in results] return final_results