Deception Detection using Crime Analytics: Leveraging Deep Learning for Safer Digital Spaces

View at publisher Summarise with VIRAT
Authors: Jasveer Saini, Aarya Dave, Yukta Makwana, Medha Asurlekar
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Today, with social media being widespread, it is challenging for law enforcement to spot the ways people hide the truth before criminal acts. It attempts to resolve these issues by proposing a model that links deception detection with crime analytics. With this method, we use Machine Learning (Random Forest) and Deep Learning (LSTM, ConvBiLSTM and RNN) together with Natural Language Processing to analyze text, sentiment and what’s included in social media posts. After collecting the data, TF-IDF and Word2Vec techniques are used to better sort through the information and classify things as genuine or fake. The produced results reveal that the ConvBiLSTM model performs best in this area with an accuracy of 91.2%. Also, when deception markers were added to crime forecasting tools, the predictions became 13–15% more accurate, proving their usefulness. As a result of this research, misinformation can be spotted and criminals predicted more easily which helps support both digital forensics and law enforcement systems.

Fields of Study & Subjects

Misinformation and Its Impacts Deception detection and forensic psychology Authorship Attribution and Profiling

Publication details

Journal / source
CRC Press eBooks
Pages
34–40
Publication date
25 March 2026
Language
EN
Document type
Book chapter