– Spectral Analysis of ST414 Time Series Data Lecture 7

advertisement
ST414 – Spectral Analysis of
Time Series Data
Lecture 7
20 February 2014
Last Time
• Asymptotics for the spectral density matrix
• Statistical inference for coherence
analyses
• Estimating VAR parameters
• Causality
• The VAR spectral density matrix
2
Today’s Objectives
• VAR spectral density matrix example
• Linear filters
• Tapering
3
The VAR Spectrum
Recall the spectral density for an AR(p) is
𝜎2
𝑓 πœ” =
,
2
|Φ(exp −𝑖2πœ‹πœ” )|
where
𝑝
πœ™π‘— 𝑧 𝑗
Φ π‘§ =1−
𝑗=1
4
The VAR Spectrum
The spectral density matrix for a VAR(p) is
𝑓 πœ” = Φ(exp −𝑖2πœ‹πœ” )−1 Σ[Φ(exp −𝑖2πœ‹πœ” )∗ ]−1 ,
where
𝑝
Φ𝑗 𝑧 𝑗
Φ π‘§ = 𝐼𝑑 −
𝑗=1
5
Sales Example
6
Sales Example
7
Sales Example
VAR(2) for Sales/Lead data:
0.280 − 0.730
Φ1 =
0.028 − 0.516
0.205 − 2.177
Φ2 =
−0.011 − 0.153
1.431 − 0.022
Σ=
−0.022
0.077
8
Sales Example
Causality tests:
𝐻0 : Sales does not Granger-cause Lead
p = 0.267
𝐻0 : Lead does not Granger-cause Sales
p = 1.806 x 10-8
𝐻0 : No instantaneous causality
p = 0.422
9
Sales Example
10
Sales Example
11
Sales Example
12
Seasonality Example
Suppose X(t) is weakly stationary, sampled
monthly. Consider the differenced series Y(t)
= X(t)-X(t-12). What is the spectrum for Y(t)?
13
Seasonality Example
14
Seasonality Example
15
Seasonality Example
16
Seasonality Example
17
Seasonality Example
18
Differencing Example
This time, consider the first difference:
Y(t) = X(t) – X(t-1)
What is the spectrum for Y(t)?
19
Differencing Example
20
Differencing Example
21
Moving Average Example
Consider now a moving average X(t):
1
π‘Œ 𝑑 =
𝑋 𝑑−6 +𝑋 𝑑+6
24
1
+
12
5
𝑋(𝑑 − π‘Ÿ)
π‘Ÿ=−5
What is the spectrum for Y(t)?
22
Moving Average Example
23
Moving Average Example
24
Moving Average Example
25
Moving Average Example
26
Linear Filters
Let X(t) be a weakly stationary process with
spectrum 𝑓𝑋 πœ” , and let πœ“π‘—
be a
π‘—πœ–π’
sequence with
π‘—πœ–π’ |πœ“π‘— |
π‘Œ 𝑑 =
< ∞. Let
π‘—πœ–π’ πœ“π‘— 𝑋
𝑑−𝑗 .
Then the spectrum of π‘Œ 𝑑 is
π‘“π‘Œ πœ” = |A πœ” |2 𝑓𝑋 πœ” ,
where A πœ” = π‘—πœ–π’ πœ“π‘— exp(−𝑖2πœ‹πœ”π‘—).
27
Linear Filters
Let X(t) be a weakly stationary P-variate process
with spectral density matrix 𝑓𝑋 πœ” , and let Ψ𝑗
π‘—πœ–π’
be a sequence of QxP matrices with π‘—πœ–π’ ||Ψ𝑗 || <
∞, where || βˆ™ || is any matrix norm. Let
π‘Œ 𝑑 = π‘—πœ–π’ Ψ𝑗 𝑋 𝑑 − 𝑗 .
Then the QxQ spectral density matrix of π‘Œ 𝑑 is
π‘“π‘Œ πœ” = A πœ” 𝑓𝑋 πœ” A πœ” ∗ ,
where A πœ” = π‘—πœ–π’ Ψ𝑗 exp(−𝑖2πœ‹πœ”π‘—).
28
Tapering
If X(t) is zero-mean weakly stationary with
spectral density 𝑓𝑋 πœ” , one can show that
0.5
𝐸 𝐼 πœ”π‘—
=
π‘Šπ‘‡ (πœ”π‘— − πœ”) 𝑓𝑋 πœ” π‘‘πœ”,
−0.5
where π‘Šπ‘‡ (πœ”) is the Fejer kernel.
29
Tapering
Let h(t) be a data taper, with DFT
𝑇
𝐻𝑇 πœ” = 𝑇 −1/2
β„Ž 𝑑 exp(−𝑖2πœ‹πœ”π‘‘) .
𝑑=1
Define π‘Œ 𝑑 = β„Ž 𝑑 𝑋(𝑑). Then
0.5
𝐸 πΌπ‘Œ πœ”π‘—
=
π‘Šπ‘‡ (πœ”π‘— − πœ”) 𝑓𝑋 πœ” π‘‘πœ”,
−0.5
where π‘Šπ‘‡ πœ” = |𝐻𝑇 πœ” |2 .
30
Example
31
Shumway & Stoffer, 2004
Example
32
Example
33
Example
34
Download