Predictive Modelling Using Museum Data

Abstract

Museums face challenges in maintaining and preserving their vast collections, particularly when identifying artworks that require restoration and detecting potential forgeries. This project leverages machine learning models to enhance museum collection management. Using data from a museum collection, Random Forest and Isolation Forest algorithms predict restoration needs and detect forgeries, respectively. The results show high accuracy in restoration prediction, with Age at Acquisition being the most significant feature. Forgery detection flagged 1,303 potential cases, providing museums with valuable insights for further investigation. This approach streamlines operational processes and ensures the long-term preservation and authenticity of art collections.