On this page
Research Article | Open Access
Volume 14 2022 | None
Crime Hotspot Analysis Using Six Layered Deep Recurrent Neural Networks
Dr.S.Meenakshi Sundaram, Dr. P. Sivakumar, Dr. M. Parthiban, Dr.U.Revathy
Pages: 2226-2237
Abstract
Crime is a prominent and alarming feature of our society. Every day, a large number of crimes are attempted, making ordinary citizens' lives difficult. As a result, it is critical to prevent the crime from occurring. The capability to anticipate future crimes can aid law enforcement in preventing them before they happen. From a deliberatepoint of view, the capacity to predict every crime based on time, place, and other factors can assist law enforcement in delivering significant information. However, because crimes are increasing at an alarming rate, precisely anticipating the crime is a challenging task. As a result, crime forecast and investigation approaches are important for identifying and preventing crimes in the future. Many academics have recently done studies applying various machine learning approaches and specific inputs to predict crimes. The existing model uses different algorithms like KNN, SVM, and Naïve Bayes which obtains low accuracy. The proposed methodology will predict the crime rate based on the given location using Recurrent Neural Networks which consists of six dense layers. The datasets Vancouver and googletrends used in this project are taken from Kaggle having 480725 and 185 tuples respectively. The proposed methodology got the 89% accuracy for the prediction of crime rate.
Keywords
DeepLearning, Crime Hotspots, Recurrent Neural Networks, Vancouver, Latitude, Longitude.
PDF
132
Views
13
Downloads