Multi-kernel Regression For Graph Signal Processing
This work addresses signal processing on graphs, which is incremental as it builds on existing kernel methods by incorporating multi-kernel techniques for potentially better performance in specific applications.
The authors tackled the problem of graph signal processing by developing a multi-kernel regression method that assumes target signals are smooth over a graph, resulting in improved efficiency over standard kernel approaches as shown in simulations with real-world data.
We develop a multi-kernel based regression method for graph signal processing where the target signal is assumed to be smooth over a graph. In multi-kernel regression, an effective kernel function is expressed as a linear combination of many basis kernel functions. We estimate the linear weights to learn the effective kernel function by appropriate regularization based on graph smoothness. We show that the resulting optimization problem is shown to be convex and pro- pose an accelerated projected gradient descent based solution. Simulation results using real-world graph signals show efficiency of the multi-kernel based approach over a standard kernel based approach.