Geospatial Discriminative Patterns with Principle Direction for Effective Crime Detection

Identifying and mapping crime have a propensity to cluster geographically. This kind of grouping led to the wise practice of crime and hotspot analysis, which helps to analyze, identify and visualize crime. The process of crime detection has two different issues, which are accuracy and mapping of crimes. The proposed system provides a visual and graphical representation of crimes using machine learning and data mining approaches. The result from the accurately identified crime reports can be more beneficial to the public. Several existing system used mapping methods for hotspot. But those existing approached failed to map crime locations effectively. So the proposed work created a framework for crime detection with visual mapping using data mining approaches. The proposed system used support finding, Geospatial discriminative patterns to gain the significant difference between the normal class and crime class. The data should be perfectly matched with the class accurately, so the system uses Genetic approach for effective fitness finding of class. The system uses both real world dataset and user constructed dataset for evaluation.