A missing temperature data estimation method using graph Fourier transform

In this paper, a missing temperature data estimation method using graph Fourier transform (GFT) is presented. First, the GFT based on graph Laplacian matrix is briefly reviewed. Then, the spectral analysis of the temperature data is provided by using GFT to show that its bandwidth is approximately limited. Based on this band-limited property, a missing temperature data estimation approach is proposed. Finally, the real temperature data measured in ten cities of south Taiwan are employed to demonstrate the usefulness of the proposed graph Fourier transform method.