Type: id
Description: unique identifier for each pump installation
Type: id
Description: date at location of pump installation
Type: feature
Description: number of events recorded
Type: feature
Description: indicator number of events < 10
Type: feature
Description: log-fold difference in number of events relative to expected defined using k-means clustering
Type: feature
Description: log-fold difference in number of events relative to expected defined as mean
Type: feature
Description: log-fold difference in number of events relative to expected defined using k-means clustering per day of week
Type: feature
Description: log-fold difference in number of events relative to expected defined as mean per day of week
Type: feature
Description: total duration of all pumping events
Type: feature
Description: log-fold difference in pump function relative to stl regression with weekly seasonality
Type: feature
Description: log-fold difference in pump function relative to expected defined as mean per day of week
Type: feature
Description: log-fold difference in flow rate relative to stl regression with weekly seasonality
Type: feature
Description: log-fold difference in number of events relative to stl regression with weekly seasonality
Type: outcome
Description: indicator that the pump will fail in the next seven days (including current date)
Type: prediction
Description: predicted probability of forecasted failure using SuperLearner (cross-validated)
Type: outcome
Description: indicator that the pump is currently failed
Type: prediction
Description: predicted probability of currently failure using SuperLearner (cross-validated)
Type: prediction
Description: predicted probability of forecasted failure using GLM (cross-validated)
Type: prediction
Description: predicted probability of currently failure using GLM (cross-validated)