Abstract: In recent decades,when analysis and process two modern financial issues of "risk management and effectiveness of the optimization" in the international community,the financial time series model and the measurement of the growing corresponding play a very important role.The linear time-series models Have been known such as the AR(p),MA(q),ARMA(p,q).There are a lot of classical methods such as the least-squares method and the maximum likelihood method can estimate parameters of many models except missing data in the middle,when the above methods will be powerless.This article will discuss in detail the data deletion ARMA(1,1) model,that is,the parameter estimation methods of Zt=αZ_t-1+ε_t-βε_t-1.

Key words: missing data, ARMA(1,1) model, likelihood function, EM algorithm