def relatives(list_name, data, relatives_num):
relatives_list = [] # 药物亲友团
length = data.shape[0]
for item in list_name:
list_ = []
for i in range(length):
words = data['药物'][i]
if item in words:
list_.append(words)
relatives_list.append(list_)
return utils.cut_by_num(relatives_list, relatives_num)
def cluster_main2(relatives_list, list_name):
list_qyt = duplicate_removal(relatives_list, list_name)
# 使用数字代替列表中的项
list_num = utils.word_2_num(list_name, list_qyt)
for group_num in range(3, 9):
new_list = utils.cut_by_num(list_num, group_num)
list_num2 = del_by_correlation(new_list)
reWord = utils.num_2_word(list_name, list_num2)
# 创建二元组
doubleSet = create_double_set(list_num2)
max_num, bestSet = merge_loop(doubleSet, list_name,
'data/group' + str(group_num) + '.csv')
# 信息利用率
print(max_num, '/', group_num, '=', max_num / group_num)
def relatives_2(list_name, data, relatives_num):
"""
根据互信息得到每项的亲友团
:param list_name:所有词的list
:param data:dataFrame，{组合，关联度系数}
:param relatives_num:限制亲友团个数
:return:[[]] 所有项的亲友团
"""
relatives_list = [[] for i in range(len(list_name))]
length = data.shape[0]
for i in range(length):
words = data['组合'][i]
# words = data['药物'][i]
pre_index = list_name.index(words[0])
relatives_list[pre_index].append(words)
suf_index = list_name.index(words[1])
relatives_list[suf_index].append(words)
return utils.cut_by_num(relatives_list, relatives_num)