Prohibitive Analysis and Futuristic Crime Prediction

Abstract

Crimes are happening daily without any limits. They must be reduced in order to make the world a better place to live. This paper aims at predicting the future crimes which are going to happen and prevent them from even happening. There are two categories in which a crime may happen. First category is that he/she may be an existing criminal. In this case the rules based prediction model may be implied in order to predict the crime. Experimental results prove that areas which are prone to a specific crime are more likely of a repetitive occurrence of the same crime. In this approach Decision Trees and Naive Bayes are used at various stages of classification through the Pandas, Numpy and Scipy frameworks. After classification and regressive fine tuning the knowledge of a specific crime is predicted. Second category is that a new person may commit a crime. In this approach Apriori algorithm may be implemented to find out the hidden patterns among different types of crimes. These patterns serve as the basis for further prediction classifier. Furthermore, the Pictorial representation of the criminal is analyzed and authenticated through LabVIEW, since the system analyses and monitors real time CCTV footages and timely response helps in reducing the Mean Square Error.