Data
subject:: Data Science Methods for Large Scale Graphs parent:: Graph Signals and Graph Signal Processing theme:: math notes
Spectral representation of a graph convolution
Spectral Representation of a Convolutional Graph Filter
in the spectral/frequency domain, we have this representation is completely defined by the polynomial
Frequency Response
The spectral representation/frequency response of is
Spectrum Representation as Sampling from a Polynomial
Because the spectral representation of a graph filter is independent from the graph and the spectral graph filter operates on a signal pointwise, this tells us
- The polynomial is fixed once we find the coefficients .
- We can think of as sampling values along the curve .
Thus, changing the shift operator corresponds to a different set of points sampled from this polynomial.
![]()
Spectral Representation of a Graph Signal
The graph fourier transform of the filtered signal is given as
\hat{y} = V^* y &= V^* \sum_{k=0}^{K-1}h_{k}S^{k} x \\ &= V^* \sum_{k=0}^{K-1}h_{k}S^{k} (V \hat{x}) \\ (*) &= \sum_{k=0}^{K-1}h_{k} V^* (V \Lambda V^*)^k V \hat{x} \\ &= \sum_{k=0}^{K-1}h_{k} \Lambda^k \hat{x} \end{aligned}$$ For $(*)$, recall the definition of the [[Concept Wiki/inverse graph fourier transform]].
We can see then that the GFT of the filter and the operator have the same polynomial relationship, but in the spectral domain.
Notes
Mentions
TABLE
FROM [[]]
FLATTEN choice(contains(artist, this.file.link), 1, "") + choice(contains(author, this.file.link), 1, "") + choice(contains(director, this.file.link), 1, "") + choice(contains(source, this.file.link), 1, "") as direct_source
WHERE !direct_source