ABSTRACT : Preventive measures are always better than curative ones. The same is true for crimes as well. The Crime Analysis uses mathematics, predictive modeling and predictive analysis to help law enforcement in targeting potential criminal and antisocial activities. Studying, observing and analyzing the patterns formed in crimes are used in various countries and organizations. Crime is dynamic in nature but still we can find patterns in a crime which will help the authoritative and concerned organizations to find areas that are less affected by a crime and those with a high rate in a particular crime. This research aims at providing people a thought on how crime patterns can be analyzed and can help in creating a crime free neighborhood with the help of supervised learning. K-nearest neighbor algorithm was used to find locations that are vulnerable to the classified crime. As of now, we have taken six classes of crime: Robbery, Accident, Gambling, Violence, Kidnapping, and Murder. For this research the datasets were selected from government websites, which were pre processed and used to find patterns in the various classes of crime occurring in different states according to the jurisdiction of the country.

KEYWORDS- Crime Analysis, Predictive Modelling, K-nearest neighbor, Supervised Learning

| DOI: 10.17148/IARJSET.2021.81135