Hi All,
Apologies if this is confusing, but I’ll try my best.
I have a set of time series data that includes geographic location with the dependent variable being the presence and location of an event type. Imagine, for example, that I am able to measure lava flows at many points over a large area every 5 minutes and I want to predict whether any eruption has occurred and where and when they occurred. Is this a classification problem? How do I represent the dependent variable? In most five minute windows, no eruption has occurred. To reiterate, I want to find the time and location of eruptions. Can anyone give me some guidance on structuring the data?
to make this clearer, here’s what my raw data looks like:
dataset 1 (independent variables):
time window 1, sensor a, sensor a coordinates, sensor a reading
time window 1, sensor b, sensor b coordinates, sensor b reading
time window 2, sensor a, sensor a coordinates, sensor a reading
time window 2, sensor b, sensor b coordinates, sensor b reading
dataset 2 (for labels):
time window 1, eruption a occurred, eruption coordinates
time window 1, eruption b occurred, eruption coordinates
time window 2, eruption didn’t occur, no eruption coordinates