I am looking for best approach for modeling inter arrival times for series of events. I have dataset that present user file creation pattern in a period of one year. I model this number of files per day using negative binomial distribution /I generated dataset using paramteres with R and compared with emipiric (colected) data. using p-value. and MLE method.
In most of cases inter arrival is measured as events/hours, but in my cases you there is some activity durring certain hours durring a day and on other hours there is no activity. Any ideas how to model this inter arrival time? Should I cluster data based on num of files per day and find inter arrival time for day with one file creation, two files creation and so on ... Any ideas are welcome.
- Badges
- Science topic
- Similar topics
- Correction