Additional Info

• File: Link
• Published in: Book of short papers SIS 2020
• Year: 2020
• Abstract: In this paper we analyze two different approaches for modeling dependent count data with long-memory. The first model we consider explicitly takes into account the integer nature of data and the long-range correlation, while the second model is a count-data long-memory model where the distribution of the current observation is specified conditionally upon past observations. We compare these two different models by looking at their estimation and forecasting performances.