Hello,

I am trying to model chat room activity. Specifically I want
to generate times of people joining the chat and times of
people exiting the chat. I also want to model the change of
the rate of people joining over time, so that I can model
transitions to/from peak chatting hours.

So far I am doing the following:

-simulation goes from time Ts to time Te

-At time Ts the first person joins the chat.

-I define a function that defines the mean of the
duration between people joining over time.

-I select the time between new people joining a chat room
using a random sample from an exponential distribution
with the specified mean

-I vary the duration that each person chats by
using random sample from an exponential distribution
with a constant specified mean


Does this sound reasonable?

Any advice would be greatly apreciated.

Jacob.Copenha...@gmail.com> wrote in message

news:1174509474.318712.186980@p15g2000hsd.googlegroups.com...





Hello,

I am trying to model chat room activity. Specifically I want
to generate times of people joining the chat and times of
people exiting the chat. I also want to model the change of
the rate of people joining over time, so that I can model
transitions to/from peak chatting hours.

So far I am doing the following:

-simulation goes from time Ts to time Te

-At time Ts the first person joins the chat.

-I define a function that defines the mean of the
duration between people joining over time.

-I select the time between new people joining a chat room
using a random sample from an exponential distribution
with the specified mean

-I vary the duration that each person chats by
using random sample from an exponential distribution
with a constant specified mean

Does this sound reasonable?

Any advice would be greatly apreciated.

Have you got any real data of activity that you can look at, on which you
can base your simulation/model?

Nick- Hide quoted text -

- Show quoted text -