Question $9/$Vraag $9$

$($a$)$ Define the concept of an analytics base table $($ABT$),$

and list three of its key features.

$($b$)$ A dataset has $a=3,b=2,c=4,$ but a value for $d$ is

missing. By imputation, what would a reasonable

guess for $d$ be$?$

$($c$)$ For a feature in a data set that is expected to have

a normal distribution, the mean is significantly

higher than the median. What does this tell us

about the values of possible outliers?

$($d$)$ You received a large data set of unknown ordering.

If you want to analyse properties of a subset of the

data, should you use top sampling or random sam$-$

pling? Give a reason for your answer.

$($a$)$ Definieer die konsep van $\text{'}$n analitika$-$basistabel, en

lys drie van die kerneienskappe daarvan.

$($b$)\text{'}$n Datastel het $a=3,b=2,c=4,$ maar $\text{'}$n waarde

vir $d$ ontbreek. Deur imputasie, wat sou $\text{'}$n redelike

afskatting van $d$ wees?

$($c$)$ Vir $\text{'}$n kenmerk in $\text{'}$n datastel wat $\text{'}$n normaalver$-$

spreiding behoort te h$,$ is die gemiddeld aansien$-$

lik ho$$r as die mediaan. Wat s$$ dit vir ons oor die

waardes van moontlike uitskieters?

$($d$)$ U het $\text{'}$n datastel ontvang met onbekende volgorde.

Indien $u\text{'}n$ analise wil doen op $\text{'}n$ substel van die

data, moet $u$ topmonstering of lukrake monstering

gebruik? Gee $\text{'}n$ rede vir $u$ antwoord.

$[2]$