Empirical Analysis of Prediction of Crime and Forecasting With IGWO-SVM Approach

In order to understand the nuances of the crime data from recent years, in-depth analysis is necessary. Involved parties in law enforcement will benefit from this procedure in terms of making arrests and guiding efforts to reduce crime. Strategically or tactically, it benefits from having a better understanding of where, when, and how crimes will occur in the future. However, given the rising incidence of crime in the modern world, developing an effective method for predicting future criminal behavior is a difficult issue. As a result, it is crucial to develop a strategy for predicting criminal activity in order to prevent further crimes and lower crime rates. Researchers are currently attempting to predict criminal behavior based on a variety of factors. SVM, CNN and ANN are just few of the prediction approaches that can be used to assess a model's accuracy. Although their studies have yielded encouraging findings, there are still major blind spots that preclude an accurate prediction of crime hubs. There have been many earlier publications of academic papers on this subject. So, the results are compiled and conducted a comprehensive assessment of all of them in this report. Proposed goal is to catalog the various forms the crime prediction method has taken and the ways in which it might be improved in the future. Preprocessing, feature extraction, and model training are the three main components of the proposed method. Preprocessing includes activities including data cleansing, integration, transformation, reduction, and discretization. It uses principal component analysis to glean features. The model is trained with IGWO-SVM. When compared to two existing models, the proposed model performs exceptionally well.