Question $3$: Distribution network topology identification $($a classification problem$)$

Establishing a Low Voltage $($LV$)$ distribution network topology provides valuable insight for

DNSPs for outage management, network planning and phase balancing. The growth in smart

meter installations at distribution level across the world provides opportunity for data$-$driven

detection of LV Network topology. The topology identification problem is a classification task,

where households with unknown connecting transfer are classified into the group connected

to one of the target transformers in the area. You will be given voltage time series for a number

of customers and the transformer they are connected. You need to hide the transformer data

for some customers $($test data$)$ and use the rest of the data $($training data$)$ to train the model.

The expectation is that you compare performance of different classifiers including those

discussed in the course.