Network-Based Transfer Learning Helps Improve Short-Term Crime Prediction Accuracy

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wu, Jiahui, Frias-Martinez, Vanessa
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910485938438144
author Wu, Jiahui
Frias-Martinez, Vanessa
author_facet Wu, Jiahui
Frias-Martinez, Vanessa
contents Deep learning architectures enhanced with human mobility data have been shown to improve the accuracy of short-term crime prediction models trained with historical crime data. However, human mobility data may be scarce in some regions, negatively impacting the correct training of these models. To address this issue, we propose a novel transfer learning framework for short-term crime prediction models, whereby weights from the deep learning crime prediction models trained in source regions with plenty of mobility data are transferred to target regions to fine-tune their local crime prediction models and improve crime prediction accuracy. Our results show that the proposed transfer learning framework improves the F1 scores for target cities with mobility data scarcity, especially when the number of months of available mobility data is small. We also show that the F1 score improvements are pervasive across different types of crimes and diverse cities in the US.
format Preprint
id arxiv_https___arxiv_org_abs_2406_06645
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Network-Based Transfer Learning Helps Improve Short-Term Crime Prediction Accuracy
Wu, Jiahui
Frias-Martinez, Vanessa
Machine Learning
Computers and Society
Deep learning architectures enhanced with human mobility data have been shown to improve the accuracy of short-term crime prediction models trained with historical crime data. However, human mobility data may be scarce in some regions, negatively impacting the correct training of these models. To address this issue, we propose a novel transfer learning framework for short-term crime prediction models, whereby weights from the deep learning crime prediction models trained in source regions with plenty of mobility data are transferred to target regions to fine-tune their local crime prediction models and improve crime prediction accuracy. Our results show that the proposed transfer learning framework improves the F1 scores for target cities with mobility data scarcity, especially when the number of months of available mobility data is small. We also show that the F1 score improvements are pervasive across different types of crimes and diverse cities in the US.
title Network-Based Transfer Learning Helps Improve Short-Term Crime Prediction Accuracy
topic Machine Learning
Computers and Society
url https://arxiv.org/abs/2406.06645