Loading [MathJax]/jax/output/HTML-CSS/jax.js
  • 1. Invertible Representation of MA(q)
  • 2. Why Invertibility is Important
  • Summary


Class Web Page



1. Invertible Representation of MA(q)

Suppose we have MA(1) model (watch the sign!) Xt=ϵt+θ1ϵt1 This is already a causal representation.

We can rewrite the equation and get Xtθ1ϵt1=ϵt.

That means we can do the same for Xt1, and write Xt1θ1ϵt2=ϵt1 We can substitute this into ϵt1 above.

Substituting this expression of ϵt1 into the first equation, Xtθ1ϵt1=ϵt.Xtθ1(Xt1θ1ϵt2)=ϵtXtθ1Xt1θ21ϵt2=ϵt

Repeat this n times, we get Xtθ1Xt1θn11Xtn+1θn1etn=ϵt If |θ1|<1, then we can write ϵt=i=0θi1Xti This is called invertible representation.

ϵt=i=0πiXti=(π0+π1B+π2B2+π3B3+)Xt=Π(B)Xt

So for our MA(q) model Xt=Θ(B)ϵtΠ(B)Xt=ϵt

Very similar to causal representation in AR(p), we have identity, Π(z)=1Θ(z) (π0+π1z+π2z2+π3z3+)=1(1+θ1z+θ2z2++θqzq)

We can calculate coefficients by moving all the polynomials to the left and matching (1θ1zθ2z2θqzq)(π0+π1z+π2z2+π3z3+)=1

For example, if we have MA(1), (1θ1z)(π0+π1z+π2z2+π3z3+)=1

To find out the coeffeicnets πi, we can match the order of z. π0=1( coefficient without z )θ1π0+π1=0( coefficient of z)θ1π1+π2=0( coefficient of z2)θ1π2+π3=0( coefficient of z3)

π0=1π1=θ1π0π2=θ1π1π3=θ1π2

π0=1π1=θ1π2=θ21π3=θ31

Invertible Representation of MA(1)

Now we get invertible representation for MA(1), ϵt=i=0πiXti=i=0θi1Xti.

To write the today’s error (innovation) ϵt, we need infinetely many past observation Xti.

When can we do this?

Invertibility Condition

If all roots of characteristic polynomial Θ(z)=1+θ1z+θ2z2++θqzq is outside of the unit circle, then MA(q) admits a invertible representation.

Same as causal condition in AR(p).

Watch the sign! We have same sign as Φ(z) in AR(p), because we started with Brockwell’s convention; Xt=ϵt+θ1ϵt1.

We assume all MA(q) we deal with are invertible.

2. Why Invertibility is Important


To Get Residual

If we have MA(3) process, Xt=ϵtˆθ1ϵt1ˆθ2ϵt2ˆθ3ϵt3 with estimated parameter ˆθi, we want to get residuals ˆϵt, and check their randomness.

In AR(p), this was intuitive. e.g. if we had AR(2), residual was Xtˆϕ1Xt1ˆϕ2Xt2=ˆϵt

In MA(q), we must use invertible expression, ˆϵt=ˆπ0Xt+ˆπ1Xt1+ˆπ2Xt2+

And To Get Predictor

We haven’t talked about forecasting, but invertibility will be used in forecasting pretty much the same way as in the residuals.



Summary

  • MA(q) process is defiend with equation Xt=Θ(B)ϵt you can check the roots of the characteristic polynomial Θ(z)=1+θ1z+θ2z2++θqzq. If all root are outside of the (complex) unit circle, then MA(q) admits a invertible representation.

  • Inverted representation allows to write today’s error using infinite sum of past observations. ϵt=i=0Xti

  • Bcause we can’t directly observe ϵt, invertibility is importatnt in calculating residuals and forecasts.