CRIMINAL SHORT LISTING AND CRIME FORECASTING BASED ON MODUS OPERANDI

Abstract

Modus Operandi is a trait that most of the living beings have, which is also identi ed as mode of operation. Comparing several crimes performed by the same criminal will reveal a speci c Modus Operandi pattern for an individual or a group. This research suggests a model in shortlisting criminal suspects by identifying these patterns using a Speci c M.O feature set and a model derived from a machine learning approach. The model creates a number of clusters which represents a unique MO classi cation. Unlike any of the classi cations used so far this classi cation is made of a collection of single MO behaviors which were exhibited in unison. This classi cation method has its own importance since it is unique and is foolproof for changes that can happen in a single MO behavior. The model has proven to be correct with above 75% of accuracy. It validates both internal and external accuracy of the created Modus Operandi classi cations within the clusters. However, the need of prior records of a criminal within the database is a dependency in this research when identifying a criminal. As future extensions of this research a prediction model and a system implementation architecture is proposed to deploy the system with this model.