-
Nested Logit
In this case the random residual is split into two components that
cancel each
other:
In this way there is an additional hierarchy level, so that
the variable tau can be
used to set different values for each one of the groups (an inner-covariance
exists).
Four steps model – Fourth steps: Assignment
models
Classification of assignment methods:
ICT4TS – Exam summary
21
Time Series (Prof. Mellia)
Big data challenges
Fives Vs of Big Data
The knowdlege discovery process
Time series and autoregressive models
Components of a Time Series
-
Trend
-
Seasonal
variation
-
Cyclical variation
-
Irregular variation
ICT4TS – Exam summary
22
Random walk
Autocovariance Function
Gives the linear dependence between two points at different time.
Autocorrelation Function
Partial
autocorrelation
ARMA models
Autoregressive moving average models.
ARIMA models
Autoregressive Integrated moving average models.
Xt is ARIMA(p,q,d) if:
Is ARMA(p,q)