Table 1.
Baseline models.
Fig 1.
Comparing probability mass functions and hazard rates of geometric and negative binomial distributions.
Table 2.
Discrete time models.
Table 3.
Continuous time models.
Fig 2.
Comparing forecasts on a perfectly periodic, synthetic data set.
The vertical red line denotes the start of the forecast period.
Fig 3.
Comparing the forecasts of different IDF approaches on a synthetic data set with alternating periods.
Table 4.
Experiments with synthetic data sets.
Table 5.
Summary statistics of data sets.
Table 6.
Experiment results for discrete-time models with Car Parts, Auto, and RAF data sets.
Table 7.
Experiment results for discrete-time models with UCI and M5 sets.
Table 8.
Ratios of model families to the Static G-Po baseline model.
Table 9.
Experiment results for continuous-time models.