def test_one_user_by_matrix():
'''
give one fixed user , recommend result,by matrix
:return:
'''
user = "******"
alpha = 0.8
graph = read.get_graph_from_data('ratings.csv')
pr_recom_resuit = personal_rank_matrix(graph, user, alpha, 100)
return pr_recom_resuit
def get_one_recom():
'''
give one fixed user , recommend result
:return:
'''
user = "******"
alpha = 0.8
graph = read.get_graph_from_data('ratings.csv')
iter_num = 100
pr_recom_result = personal_rank(graph, user, alpha, iter_num, 100)
return pr_recom_result
'''
def get_one_user_recom():
user = '******'
alpha = 0.8 # 0.7
graph = read.get_graph_from_data('../data/ratings.txt')
iter_num = 100 # 15
recom_result = personal_rank(graph,user,alpha,iter_num)
[print(i) for i in recom_result]
print("========================================")
# 解析，便于理解
item_info = read.get_item_info('../data/movies.txt')
# 将用户现在感兴趣的item打印出来
for itemid in graph[user]:
pure_itemid = itemid.split("_")[1]
print(item_info[pure_itemid])
print("=================推荐结果=================")
# 将用户可能感兴趣的结果打印出来
for itemid in recom_result:
pure_itemid = itemid.split("_")[1]
print(item_info[pure_itemid])
print(recom_result[itemid])
def mat_all_point(m_mat, vertex, alpha):
"""
get E-alpha*m_mat.T
Args:
m_mat:
vertex: total item and user point
alpha: the prob for random walking
Return:
a sparse
"""
total_len = len(vertex)
row = []
col = []
data = []
for index in range(total_len):
row.append(index)
col.append(index)
data.append(1)
row = np.array(row)
col = np.array(col)
data = np.array(data)
eye_t = coo_matrix((data, (row, col)), shape=(total_len, total_len))
return eye_t.tocsr() - alpha * m_mat.tocsr().transpose()
if __name__ == "__main__":
graph = read.get_graph_from_data("../data/log.txt")
m, vertex, address_dict = graph_to_m(graph)
mat_all_point(m, vertex, 0.8)
def get_one_user_by_mat(user):
alpha = 0.6
graph, graph_count = read.get_graph_from_data(r"..\data\企业交易数据all.txt")
recom_result = personal_rank_mat(graph, graph_count, user, alpha)
return recom_result
alpha:the prob for random walking随机游走概率
return:
sparse matrix
"""
# np.eys()初始化单位矩阵易超内容
total_len = len(vertex)
row = []
col = []
data = []
for index in range(total_len):
row.append(index)
col.append(index)
data.append(1)
row = np.array(row)
col = np.array(col)
data = np.array(data)
eye_t = coo_matrix((data, (row, col)), shape=(total_len, total_len))
print(eye_t.shape)
return eye_t.tocsr() - alpha * m_mat.tocsr().transpose()
if __name__ == "__main__":
graph = read.get_graph_from_data('../data/ratings.txt')
m, vertex, address_dict = graph_to_m(graph)
print(len(address_dict))
print(m.todense())
print(m.todense().shape)
res_m = mat_all_point(m, vertex, 0.8)
print(res_m.shape)
print(res_m)
def get_one_user_by_mat():
user = '******'
alpha = 0.8
graph = read.get_graph_from_data('../data/ratings.txt')
recom_result = personal_rank_mat(graph,user,alpha,10)
return recom_result
def mat_all_point(m_matrix, vertex, alpha):
'''
矩阵算法personal_rank的公式
:param m_matrix:
:param vertex: 所有(item+user)顶点
:param alpha: 随机游走的概率
:return: 矩阵
'''
# 初始化单位矩阵（如果使用numpy创建，容易超内存）
total_len = len(vertex)
row = []
col = []
data = []
for index in range(total_len):
row.append(index)
col.append(index)
data.append(1)
row = np.array(row)
col = np.array(col)
data = np.array(data)
eye_t = coo_matrix((data, (row, col)), shape=(total_len, total_len))
return eye_t.tocsr() - alpha * m_matrix.tocsr().transpose()
if __name__ == '__main__':
graph = read.get_graph_from_data('log')
m, vertex, address_dict = graph_to_m(graph)
print(address_dict)
def mat_all_point(m_mat, vertex, alpha):
"""
get E-alpha*m_mat.T
m_mat: m matrix
vertex: all point
alpha: the prob of walk
return:
a sparse matrix
"""
total_len = len(vertex)
row = []
col = []
data = []
for index in range(total_len):
row.append(index)
col.append(index)
data.append(1)
row = np.array(row)
col = np.array(col)
data = np.array(data)
eye_t = coo_matrix((data, (row, col)), shape=(total_len, total_len))
return eye_t.tocsr() - alpha * m_mat.tocsr().transpose()
if __name__ == "__main__":
graph = read.get_graph_from_data(os.path.dirname(os.path.dirname(os.path.dirname(os.path.dirname(__file__)))) + "\\recommender\\data\\log.txt")
m, vertex, address_dict = graph_to_m(graph)
# print(address_dict)
# print(m.toarray())
print(mat_all_point(m, vertex, 0.8))