A novel approach for intelligent crime pattern discovery and prediction

Crime analysis plays a major part in crime prevention and safety of people in a country. The paper focuses on state based frequent crime pattern knowledge discovery and prevention. Our work concentrates on Finding frequent crimes state wise using FP Max a bottom up approach which uses linked lists for reduction of space complexity. The generated frequent crime sets of state will be undergone through knowledge discovery process. Correlation between the crime types is done to find the weightage factor of crime types to find crime intensity point. The crime intensity point of 29 states and 7 union territories are calculated according to the weightages derived from the correlation analysis. Later we classified the states as most dangerous, dangerous, moderate or safe based on their crime intensity point using Random forests classification technique. Prediction of crime intensity point for the state based on the responsible factors that contribute to a crime.