Questions tagged [spectral-graph-theory]

For questions related to the study of properties of a graph in relationship to the spectral properties of some associated matrix.

In mathematics, spectral graph theory is the study of properties of a graph in relationship to the characteristic polynomial, eigenvalues, and eigenvectors of matrices associated to the graph, such as its adjacency matrix or Laplacian matrix.

An undirected graph has a symmetric adjacency matrix and therefore has real eigenvalues (the multiset of which is called the graph's spectrum) and a complete set of orthonormal eigenvectors.

While the adjacency matrix depends on the vertex labeling, its spectrum is graph invariant.

531 questions
80
votes
6 answers

Motivation for spectral graph theory.

Why do we care about eigenvalues of graphs? Of course, any novel question in mathematics is interesting, but there is an entire discipline of mathematics devoted to studying these eigenvalues, so they must be important. I always assumed that…
45
votes
1 answer

What do the eigenvectors of an adjacency matrix tell us?

The principal eigenvector of the adjacency matrix of a graph gives us some notion of vertex centrality. What do the second, third, etc. eigenvectors tell us? Motivation: A standard information retrieval technique (LSI) uses a truncated SVD as a…
23
votes
1 answer

What is the intuition behind / How can we interpret the eigenvalues and eigenvectors of Euclidean Distance Matrices?

Given a set of points $x_1,x_2,\dots,x_m$ in the euclidean space $\mathbb{R}^n$, we can form a $m\times m$ Euclidean Distance Matrix $D$ where $D_{ij}={\|x_i-x_j\|}^2$. We know a little bit about these matrices like: It is symmetric. Its Trace is…
20
votes
4 answers

On the invertibility of the adjacency matrix of a graph

Which are the sufficient and necessary conditions for an undirected graph with no self edges (i.e. no loop of length $1$) to have an invertible adjacency matrix? In this case, the adjacency matrix is symmetric (i.e. $A = A^\top$). Moreover, all the…
18
votes
1 answer

What can we say about the graph when many eigenvalues of the Laplacian are equal to 1?

The Laplacian of the graph has all the eigenvalues real and non-negative, the smallest being 0. I have a graph where the second smallest eigenvalue (the so called algebraic connectivity) is equal to $1$. In fact, the multiplicity of this eigenvalue…
16
votes
1 answer

What does the minimal eigenvalue of a graph say about the graph's connectivity?

I'm reading Fan Chung's Spectral Graph Theory, and I'm now in chapter 2. There, Chung proves Cheeger's inequality, which is that $2h_G \geq \lambda_1 > h_G^2/2$ for any graph $G$. To me, this inequality raises the question of a physical…
15
votes
3 answers

Why Laplacian Matrix need normalization and how come the sqrt of Degree Matrix?

Why Laplacian matrix needs normalization and how come the sqrt-power of degree matrix? The symmetric normalized Laplacian matrix is defined as $$\ L = D^{1/2}AD^{-1/2}$$ where L is Laplacian matrix, A is adjacent matrix. Element $A_{ij}$ represents…
John Buger
  • 153
  • 1
  • 1
  • 4
14
votes
1 answer

Theoretical link between the graph diffusion/heat kernel and spectral clustering

The graph diffusion kernel of a graph is the exponential of its Laplacian $\exp(-\beta L)$ (or a similar expression depending on how you define the kernel). If you have labels on some vertices, you can get labels on the rest of the vertices by a…
14
votes
3 answers

Can I find the connected components of a graph using matrix operations on the graph's adjacency matrix?

If I have an adjacency matrix for a graph, can I do a series of matrix operations on the adjacency matrix to find the connected components of the graph?
13
votes
1 answer

Spielman's proof of graph connectivity

I use Spielman's lectures on course Spectral Graph Theory I have few question regarding Lecture 2. The Laplacian, especially Lemma 2.3.1 (Graph connectivity). Please, help me to make it a little bit clearer. Lemma 2.3.1. Let $G=(V,E)$ be a graph,…
13
votes
1 answer

Intuitive interpretation of the adjacency matrix as a linear operator.

Naturally we can describe graphs via tables of "yes there is an edge" or "no there is not" between each pair of vertices, so the definition of an adjacency matrix is easily understood. Thinking of these tables as matrices, however, adds structure -…
12
votes
3 answers

Geometric intuition of graph Laplacian matrices

I am reading about Laplacian matrices for the first time and struggling to gain intuition as to why they are so useful. Could anyone provide insight as to the geometric significance of the Laplacian of a graph? For example, why are the eigenvectors…
12
votes
1 answer

How to read Spectral Theory of Graphs

My background is a course is Linear Algebra -Hoffman,Kunze Graph Theory-Frank Harary I am doing a coursework in Spectral Graph Theory . As I am going through it, I am searching for some applications in this topic. One application I found was…
12
votes
1 answer

Eigenstructure of discrete Laplacian on uniform grid

The discrete Laplacian of a graph is the matrix $L = D - A$ where $D$ is a diagonal matrix with $d_{ii}$ being the degree of $v_i$, and $A$ is the usual adjacency matrix. Is there anything known about the eigenstructure of the discrete Laplacian of…
Suresh Venkat
  • 1,282
  • 9
  • 25
12
votes
5 answers

Spectrum of adjacency matrix of complete graph

Fooling around in matlab, I did an eigenvalue decomposition of the adjacency matrix of $K_5$. A = [0 1 1 1 1; 1 0 1 1 1; 1 1 0 1 1; 1 1 1 0 1; 1 1 1 …
1
2 3
35 36