Support Vector Machine Algorithm for Analysis of FBI Crime Data

Abstract

Cyber-incidents are a mixture of discrete instances with new illegal acts. Cybercrime incidents occur
as separate criminal offences and, according to the national crime statistics and surveys, the instances
are increasing. The existing system classifies the cybercrimes and cyber-incidents with less accuracy.
These overlapping in its result and the lack of a unique algorithm for classification are the main
drawback of the existing model.Thus, to solvethese problems, the paper gives an exclusive way of
classifying various crimes depending on the physical factors such as time and date. It gives a solution
to users to carry out an easy efficient classification outcome using a cybercrime classifier with support
vector machine (SVM). It uses the grouping of the dataset by either decision trees or random forest to
build up a model to prepare over a preparation set in order to get the most exact outcomes. It is a
modest and productive approach to group cybercrimes with the goal that the affected can identify the
kind of occurrence and follow-up correspondingly. Additionally, it examines the information and
creates various charts for the correct portrayal of the information. The above model designed to
categorize convicted criminals into low, medium and high risk of turning into recidivists helps curb
the increasing crime rates in the society, thus ensuring the welfare and well-being of its citizens. The
extensive simulation results show that the proposed method gives the outstanding classifier compared
to the state of art approaches.