1. 2019年5月31日北美考题。由于论坛里广为流传的是Mikey大神的2015年版的SBA机经,今年的SBA题库发生了较大变化,可能是与SBA出了新的教材有关。本机经将在旧机经的基础上补充新题的内容,想要旧题机经的可以从我这下载。
新考点:
1。 Studentized Residual
|SR| > 2 for data sets with a relatively small number of observations
|SR| > 3 for data sets with a relatively large number of observations.
2. 如何找到Influential observations.
􀂄 STUDENT residual
􀂄 Cook’s D
􀂄 RSTUDENT residual
􀂄 DFFITS.
3. 在Model 中,selection = Score 的意义。答案是:SCORE requests best subset selection
4.能掌握 param = ref, ref = " ".
class gender (param=ref ref='Male')
income (param=ref ref='Low');
5.
如何计算C, 这题我也没准备
Suppose there are a total of
pairs with different responses:
of them are concordant,
of them are discordant, and nt =
of them are tied.
C = (nc + 0.5nt) / total
以防万一,其他correlation rank index的算法这里。
6. EDF analysis。考哪一个语句可以求到EDF maximum devation from mean.抱歉这一块我没有准备到,请各位大神们自己参考下。EDF官网信息
7. SS model = sum of type I SS
8. Which method to split dataset when sample size is small:
􀂄 k-fold。 我选择的
􀂄 3-way partition
􀂄 不记得了
想要旧机经欢迎下载。