Crime Trend Analysis and Prediction Using Mahanolobis Distance and Dynamic Time Warping Technique

Recently, interest has been increasing in crime analytics in which time series clustering research has been playing an important role, particularly for finding useful similar trends in crime. Time series clustering technique has potential for large volume of data analysis. Multivariate time series data relating to various areas such as finance, health sector, environmental research, and crime is very useful to determine and apply data mining techniques. Analyzing multivariate time series data at different points of time helps reveal crime trends in which law enforcement agencies are interested. Police administration at state and districts level make use of the trend analysis to solve new crime cases and help them prevent future possibilities of similar kind of crime. This paper presents a new approach using dynamic time warping technique and Mahanolobis distance model to use records and statistics from to analyse crime trends and predict future crime.