Real world data sets for Activity Recognition with Ambient Sensing
The houses are real homes with multiple residents.
The residents did not follow a specific scenario. We collected real world activities from real people in real homes.
Set of activities
Set of binary binary sensors.
In total we have
2 houses with multiple residents
Full month of data per house
Over 26 millions of sensor readings
Over 3000 activity occurrences
We have ground truth labels for activities.
You can develop new sophisticated machine learning algorithms for future smart homes
evaluate your performance