Analysis of Irregular Spatial Data with Machine Learning: Classification of Building Patterns with a Graph Convolutional Neural Network (Short Paper)

Machine learning methods such as Convolutional Neural Network (CNN) are becoming an integral part of scientific research in many disciplines, the analysis of spatial data often failed to these powerful methods because of its irregularity. By using the graph Fourier transform and convolution theorem, we try to convert the convolution operation into a point-wise product in Fourier domain and build a learning architecture of graph CNN for the classification of building patterns. Experiments showed that this method has achieved outstanding results in identifying regular and irregular patterns, and has significantly improved in comparing with other methods