{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/attention-based-spatial-interpolation-for","title":"Attention-Based Spatial Interpolation for House Price Prediction","arxiv_id":null,"date":"2021-10-21","proceeding":"International Conference on Advances in Geographic Information Systems 2021 10","authors":["Darniton Viana","Luciano Barbosa"],"abstract":"Estimating the market price of a house is important for many busi-\r\nnesses such as real estate and mortgage lending companies. The\r\nprice of a house depends not only on its structural features (e.g. area\r\nand number of bedrooms) but also on the spatial context where it is\r\nlocated. In this work we estimate the price of a house based solely\r\non its structural features and the characteristics and price of its\r\nneighbors. For that, we propose a hybrid attention mechanism that\r\nweights neighbors based on their similarity to the house in terms\r\nof structural features and geographic location. For the structural\r\nfeatures, we apply an euclidean-based attention and, for the geo-\r\ngraphic location, we propose an attention layer based on a radial\r\nbasis function kernel. Those attention mechanisms are then used\r\nby a neural network regressor to predict the price of a house and to\r\ngenerate a vector representation of the house based on its implicit\r\ncontext: the house embedding, which can be used as a feature set\r\nby any regressor to perform house price prediction. We have per-\r\nformed an extensive experimental evaluation on real-world datasets\r\nthat shows that: (1) regressors using house embedding obtained\r\nthe best results on all 4 datasets, outperforming baseline models;\r\n(2) the learned house embedding improves the performance of the\r\nevaluated regressors in almost all scenarios in comparison to raw\r\nfeatures; and (3) simple regressor models such as Linear Regres-\r\nsion using house embedding achieved comparable results to more\r\ncompetitive algorithms (e.g. Random Forest and Xgboost).","url_abs":"https://dl.acm.org/doi/10.1145/3474717.3484257","url_pdf":"https://profluciano.github.io/home/pub/SIGSPATIAL21.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"attention-based-spatial-interpolation-for","repo_url":"https://github.com/darniton/ASI","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"spatial-interpolation","task_name":"Spatial Interpolation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}