Code MEM

An Ox Package for Estimating, Forecasting and Simulating Multiplicative Error Models.

  1. Le projet
  2. Disclaimer

Le projet

Multiplicative Errors Models (MEM) models are primarily concerned to model non-negatively time series. The most popular MEM type model is the Autoregressive Conditional Duration model (ACD -[Engle and Russell, 1998]) whose objective is the modeling of time between events. This model has led to the developments of a plethora of extensions, see [Pacurar, 2008] for a review.

This package intends to ease modelling of univariate Multiplicative Errors Models. Estimation, simulation and forecasting of the following models are currently supported: ACD, LOGACD, LOGACD2, PACD, EXACD, BCACD and FIACD. Generalised Autoregressive Score (GAS) models for positive series are also implemented. Available distributions for the errors are : Exponential, Generalised Gamma, Weibull and Burr.

Examples of output:

Mem output


MEM output PACD Gamma
Mem Tests PACD Gamma


This package is functional, but no warranty is given whatsoever. Any bugs, typos in the manual, or suggestions for improvements should be reported to Malick Fall (email: and will be greatly appreciated.