-
Notifications
You must be signed in to change notification settings - Fork 7
Open
Description
您好,我想问一下,关于您论文第4页的DIS Distance这一小节,最后提到
Similarly, the DIS distance between GS and NT is larger than that between LS and NT (both less than 0)
可是我用第5页的示例数据计算了一下,根据您的描述,GS对应R_7,LS对应R_8。GS和NT的DIS距离应该是小于LS和NT之间的DIS距离的呀
NT = [[1, 2, 3, 4, 12, 20, 28, 29, 30, 38, 46, 54],
[1, 2, 3, 4, 12, 20, 28, 36, 44, 52, 53, 54]]
GS = [1, 10, 19, 28, 36, 45, 54]
LS = [1, 2, 11, 20, 28, 36, 44, 52, 53, 54]
# get_both_count(a,b) 得到a,b列表中共同元素个数
s1 = (len(GS) - len(NT[0])) / (get_both_count(GS, NT[0]))
s11 = (len(GS) - len(NT[1])) / (get_both_count(GS, NT[1]))
s2 = (len(LS) - len(NT[0])) / (get_both_count(LS, NT[0]))
s21 = (len(LS) - len(NT[1])) / (get_both_count(LS, NT[1]))
print('GS\ts1:{:.4f}\ts11:{:.4f}'.format(s1, s11))
print('LS\ts2:{:.4f}\ts22:{:.4f}'.format(s2, s21))输出结果:
GS s1:-1.6667 s11:-1.2500
LS s2:-0.4000 s22:-0.2222包括您后面第10页还提到会将异常分数降序排列,前n_GS$个即为GS,剩下的n_LS个即为LS,这就是错误的吧?我看了半天不太理解,也有可能是我看漏了什么地方,希望能够解答一下。
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels