Abstract: The paper proposes a sequential monitoring scheme for detecting changes in parameter values for general time series models using pairwise likelihood. Under this scheme, a change-point is declared when the cumulative sum of the first derivatives of pairwise likelihood exceeds a certain boundary function. The scheme is shown to have asymptotically zero Type II error with a prescribed level of Type I error. With the use of pairwise likelihood, the scheme is applicable to many complicated time series models in a computationally efficient manner. For example, the scheme covers time series models involving latent processes, such as stochastic volatility models and Poisson regression models with log link function.

Key words and phrases: Composite likelihood, on-line detection, Poisson regression model, quickest detection, sequential monitoring, stochastic volatility.