An Approach to Predict the Location of Crime Using Machine Learning

Abstract

AbstractIn order to prevent crimes, it is very important to recognize the patterns of the criminal activities. If the crime patterns of different geological points of a city are known to the police and detective agencies then they can work more efficiently in order to solve the problem. The proposed methodology provides an automated technique to predict the geological location of a crime depending upon the date, time, and type of the crime from past crime behavior. Linear regression and support vector regression algorithm are used for this purpose. The proposed methodology is tested on a dataset containing the crime records of Indore city in the month of February and March 2018. The result obtained using linear regression and support vector regression algorithms is compared, and it is found that support vector regression provides better result.KeywordsCrime predictionMachine learningLinear regressionSupport vector regressionRMS error