Dear All, I am struggling to develop a model to forecast the daily expenses from a bank account. The daily time series consists (obviously) of non-negative numbers which can be zero in the days when no money is taken from the bank account. To give you an idea of the kind of series I am dealing with, please have a look at myts<-structure(c(5.5, 0, 126.93, 0, 0, 0, 0, 10, 0, 135.34, 0, 0, 0, 0, 98.21, 0, 112.38, 0, 0, 0, 0, 0, 1373.77, 151.83, 26.66, 205.5, 129.33, 172.5, 0, 10, 131.09, 0, 0, 0, 0, 0, 689, 0, 0, 0, 0, 0, 0, 60.6, 183, 98.21, 0, 0, 0, 0, 1433.79, 175.89, 0, 0, 0, 200, 134.33, 98.26, 112.21, 0, 0, 0, 0, 0, 0, 112.31, 0, 0, 0, 0, 120, 0, 350, 0, 0, 98.21, 0, 0, 0, 113.24, 0, 0, 0, 0, 15, 696.65, 321.87, 929, 210.58, 0, 0, 10), .Tsp = c(16563, 16654, 1), class = "ts") (the time origin is a bit funny, but what matters is that I have daily data). Do you know any R package to handle this kind of series? I think I am outside the domain of the ARIMA approach. I experimented with acp and tscount (to see if I could treat the series as an autoregressive Poisson series), but I did not get very far. Any suggestion is appreciated. Cheers Lorenzo