Hitachi, the conglomerate from Japan, has claimed that they have successfully developed a system that can predict crimes before they happen in real life. The system takes huge amounts of data from a variety of sources and uses the system’s Artificial Intelligence and machine learning to find patterns in crime incidents that are invisible to humans.
The system dubbed Hitachi Visualisation Predictive Crime Analytics (PCA) jogs the memory of the famous movie Minority Report. Just like the movie, the new system promises to reduce crime and enforce better policing.
The technology behind the system comes from Mark Jules and Darrin Lipscomb who were the co-founders of crime monitoring tech companies Pantascene and Avrio. Hitachi acquired these two companies last year in September.
Jules said that crime prediction is not a new idea and forecasting crime has been traditionally used to build prediction models by law enforcers. Traditionally, experience with certain variables were used to give weight to the prediction models. The model developed by Hitachi removes human bias and decides correlation between data points independent of any human interaction.
It may seem that its stupid but most criminals use social platforms for planning and boasting about their crimes but use code words for all their work. Adding data from social media can help improve the accuracy of crime prediction model by as much as 15 percent.
PCA is compatible with a whole lot of data inputs, much more than anyone could have expected. It takes the data feeds and takes a couple of weeks to determine if there are any relations in the data points. The system is so diverse that it can compensate for variables like proximity to schools, subways stations, rescue calls, gunshot reports, public places and even make note of the weather when predicting crime.
Hitachi is currently planning to trial the system in about half a dozen cities starting in October. The names of the trial cities have not been revealed yet but it has been stated that the system would initially run in the background for testing and then get trailed for active use.
via Fast Company