Extraction of characteristic substructures of room layouts focusing on rent forecast residuals

Tomonobu OZAKI
JSAI Technical Report, SIG-KBS  
The rent for rental property is determined by various factors such as area and age. In this paper, to extract interesting patterns strongly related with rent, we associate a class label to rental property based on estimated residual, and extract contrasting subgraph patterns considering the number of occurrences. Furthermore, we try to extract decision rules consisting of subgraph patterns by using a method in interpretability research of machine learning. The effectiveness of the proposed
more » ... works is evaluated in preliminary experiments using real world datasets.
doi:10.11517/jsaikbs.117.0_04 fatcat:q5fydr3xdvg7np233o2rvoa4zq