Work­shop Pa­per “A Com­par­is­on of Multi-View Learn­ing Strategies for Satel­lite Im­age-Based Real Es­tate Ap­prais­al” ac­cep­ted at AAAI 2021

 |  AllgemeinesForschung - Research

The workshop paper “A Comparison of Multi-View Learning Strategies for Satellite Image-Based Real Estate Appraisal” (by Jan-Peter Kucklick and Oliver Müller) is presented at the Workshop on Knowledge Discovery from Unstructured Data in Financial Services at AAAI 2021, which is held online at February 09th, 2021. The paper focuses on different multi-view neural network strategies to combine structured housing attributes, such as size and age, with unstructured data, such as satellite images, to increase the predictive accuracy.

The paper is available here: https://aaai-kdf.github.io/kdf2021/assets/pdfs/KDF_21_paper_12.pdf