A Model for mapping crime hotspots using neural networks: a case of Nairobi
Date
2024
Authors
Echessa, R. G.
Journal Title
Journal ISSN
Volume Title
Publisher
Strathmore University
Abstract
Since the inception of the first modernized police agency, the primary objective of police organizations has been to prevent crime. Law enforcement, police, and crime reduction agencies commonly used hotspot mapping, an analytical technique, to visually determine the locations where a crime was most prevalent. This assisted in decision-making to determine the deployment of resources in target areas. This study aimed to investigate crime mapping techniques in crime analysis and suggest ways to enhance the implementation of crime mapping in Nairobi. Beginning with a historical analysis of GIS and crime mapping, the study then moved on to a consideration of the significance of geography in dealing with crime concerns. Neural Networks and K-Means machine learning models were used, and data was collected through quantitative and qualitative means in two phases. X was utilized in the first phase to collect information from the general public and important informants. The second phase involved collecting crime hotspot coordinates using a participative Geographic Information System. The study focused on utilizing social media data and machine learning techniques, particularly the KMEANS with NN (Neural Network) model, to identify and map crime hotspots in Nairobi. By analyzing crime-related tweets and categorizing them as either positive, negative or neutral using this NN (Neural Network) model then clustering them as either high risk or low risk using K-Means, the study achieved high accuracy, precision, recall, and Fl-Score, suggesting the effectiveness of this approach for crime prediction and prevention.
Keywords: Hotspot mapping, X, Machine learning, crime, Neural Networks, K-Means
Description
Full - text thesis
Keywords
Citation
Echessa, R. G. (2024). A Model for mapping crime hotspots using neural networks: A case of Nairobi [Strathmore University]. http://hdl.handle.net/11071/15645