Question $7/$Vraag $7$

In order to apply machine learning to a particular prob$-$

lem, your team collected $12000$ weather data points of $10$

input features. Your team also collected the corresponding

power output data from a large solar photovoltaic power

system. You wish to evaluate various techniques and

learning parameters to best use future inputs to predict

system power output, given the weather conditions.

$($a$)$ Is this an example of supervised or unsupervised

learning? Motivate your answer.

$($b$)$ Clearly explain what is meant by a "training set",

"validation set" and "test set" by referring to the

function of each data set in terms of model devel$-$

opment.

$($c$)$ Indicate how you much of the data set $($as percent$-$

ages$)$ you would allocate to each of the "training

set", "validation set" and "test set". Motivate your

answer.

Om masjienleer op $\text{'}$n spesifieke probleem toe te pas, het

jou span $12000$ weer$-$datap$-$unte versamel wat $10$ intreek$-$

enmerke het. Jou span het ook die ooreenstemmende

drywings$-$uittree$-$data versamel vir $\text{'}$n groot fotovolta$-$

$$ese sonkragsentrale. Jy wil graag verskeie tegnieke en

leerparameters uitprobeer om die beste manier te vind

om toekomstige intrees te gebruik om die kragstelsel se

drywing te voorspel, gegwe die weersomstandighede.

$($a$)$ Is hierdie $\text{'}n$ voorbeeld van geleide of ongeleide

leer? Motiveer jou antwoord.

$($b$)$ Verduidelik wat bedoel word met $\text{'}$n "afrigtingstel",

"bevestigingstel" and "toetsstel" deur te verwys na

die funksie van elke datastel in terme van model

ontwikkeling.

$($c$)$ Dui aan hoeveel van die datastel $($as persentasie$)$

jy sal allokeer aan elkeen van die "afrigtingstel",

"bevestigingstel" en "toetsstel". Motiveer jou ant$-$

woord.