def test_cbr_main():
# 基于相似度的推荐
# 数据预处理
types, projs = load_dataset()
types_cates, m, cates = load_type_feature()
tf = get_type_frequency(projs, types_cates)
types_list = list(types_cates.keys())
out_str = ""
for idx in random.sample(
range(len(types_list)),
20): # range(len(types_list)):# range(20): # 取任意20个type进行推荐测试
# 获取推荐结果
recom_result = recom2(
m.getrow(idx).todense(), m, types_cates, types, tf)
print("=========target==========")
print(types[types_list[idx]])
out_str += "=========target==========\n"
out_str += "{}\n".format(types[types_list[idx]])
print("=========recom list==========")
out_str += "=========recom list==========\n"
for sim_score, type in recom_result:
print(sim_score, type)
out_str += "{} {}\n".format(sim_score, type)
print()
with open(os.path.join(
os.path.dirname(os.path.dirname(os.path.abspath(__file__))),
r"data/cb_recom_result.txt"),
"w",
encoding="utf8") as fw:
fw.write(out_str)
def gen_train_data():
"""
生成item2vec的训练数据，并写入”../data/train_data.txt“
"""
types, projs = load_dataset()
train_data_file = os.path.join(
os.path.dirname(os.path.dirname(os.path.abspath(__file__))),
r"data/train_data.txt")
produce_train_data(projs, train_data_file)
def main(preP=True, genF=True):
if preP:
types, projs = gen_type_and_proj_datasets()
else:
try:
types, projs = load_dataset()
except:
types, projs = gen_type_and_proj_datasets()
if genF:
types_cates = get_type_cate2(types)
m, cates = cate2feature(types_cates)
def test_get_train_data_main():
# 数据预处理
# preprocessed_data = load_datas(Config())
types, projs = load_dataset()
# 获取item信息
type_info = get_type_info(types, projs)
print(len(type_info))
print([type_info.get(k) for k in list(type_info.keys())[:10]])
train_data = get_train_data(projs)
print(len(train_data))
print(train_data[:50])
def test_lfm_main():
"""
test lfm model train
"""
types, projs, _ = load_dataset()
train_data = get_train_data(projs)
user_vec, type_vec = lfm_train_func(train_data, 50, 0.01, 0.1, 50)
# print(user_vec, type_vec)
type_info = get_type_info(types, projs)
for proj_key in projs:
print(proj_key)
recom_list = get_recom_result(user_vec, type_vec, proj_key)
# print(recom_list)
ana_recom_result(type_info, train_data, proj_key, recom_list)
def test_pr_main():
# 数据预处理
types, projs = load_dataset()
user = next(iter(projs.keys()))
print(user)
recom_result_base = get_one_user_recom(types, projs, user)
# graph = get_graph_from_data(data_json_list)
# m, vertex, address_dict = graph_to_m(graph)
# print(mat_all_point(m, vertex, 0.8))
recom_result_mat = get_one_user_recom_by_mat(types, projs, user)
num = 0
for elem in recom_result_base:
if elem in recom_result_mat:
num += 1
print(num)
def main():
# gen_type_and_proj_datasets(False)
load_dataset()