foreign [Music] model is modified to introduce a seasonal component right if you can call it a seasonal arima where we had our original idea of the order for Auto regression Q was for moving average and D denoted the differencing or the integration that we did right so now when we talk about seasonal arima we are introducing additional Auto regression moving moving average and differencing which correspond to the seasonal factor of your data right so this would correspond to a certain seasonality and you would Define your seasonality in terms of say if your season was one year where you could be discussing the onset of monsoon right where certain precipitation is being measured or you could talk about a certain Festival that is recurring say that's your seasonal data and your original idea of lags for the sarima model for the yes oh sorry but we had the original idea of a lack of one month right or likes were defined in in that sense right this won't be t all right so if we said we had five lags you're talking about uh T minus 1 going up to T minus five right here if we have one year as the seasonality that would essentially be defined as 12 times are uh or essentially you could say 12 times 1 right where I want to introduce one was talking about the original lag okay so that becomes if we if we're talking about One lag for the seasonality right we are talking about 12 months or one year if you're talking about two lakhs where the seasonal element is concerned we're talking about 24 months or two years so these are the orders that we would Define inside our model right we're getting this in Python from stats models dot TSA right from here we both are when we put we go for Siri Max let's try and understand where the x is coming from so our model would have surimax which talked about the time series data it would give the order which referred to are pgq values right these could be of course iterated over several values let's say we have it defined as some fixed values right and allowed that to order right so water is allotted to this you would have the subsequent values over there as a list and uh if you had the seasonal order given by Capital PDQ then we also have another parameter for the seasonal order inside our model lastly we have a exogenous component right where we could give some kind of data let's say it is data X for example all right which is referred referring to some parallel kind of Time series which is not directly related to our original time series it accounts for any kind of external effects that we are trying to incorporate into your sarima model so there we we call that as exogenous data right and introduce that as part of our model so that essentially becomes a sorry maximum all right so where we had arima earlier we are able to introduce seasonality right and also talk about bringing in external data as part of our study Max model when we are using the python modeling under statsmodel.tsa we introduce both Serima as well as surimax of course if we are talking about not having any exclamation data this by default would be ignored all right so we can operate both Serima and sarimax in the same framework