Using the Bayes Theorem in Predictive Policing
Source: https://theconversation.com/artificial-intelligence-is-used-for-predictive-policing-in-the-us-and-uk-south-africa-should-embrace-it-too-191266
In class, we learned the Bayes Theorem, a very important tool to calculate the probability of an event occurring based on some prior knowledge of conditions that may be related to that event. For example, we calculated the probability that a person has a rare disease given that he tested positive. There are many other practical uses of this method in real life, including artificial intelligence, astrophysics, finance, gambling, and weather forecasting. A recent approach combines artificial intelligence and the Bayes theorem for predictive policing, a way to predict and prevent crime in advance. As the police try to predict possible crimes, gun incidents, burglaries, and thefts can be reduced by a lot. In this way, scientists can better analyze the data and get more insights for detecting underlying patterns and relationships.
Firstly, the research builds on the Naive Bayes algorithm and uses it as a crime predictor. The researchers then married it with another algorithm called Recursive Feature Elimination, which can select the most significant features and eliminate the weaker ones, thus improving the prediction rates. The combined model is applied to an experimental dataset from Chicago with incident-level crime data, registered offenders, community concerns, and locations of police stations in the city. Compared to the original model, this new algorithm improves the predictions of the Bayes by 80%.
Still, there’re some limitations to this method. Lots of relevant data are needed to apply the algorithm. For example in South Africa, such data are hard to obtain, because of confidential issues of local police. These predictive algorithms can also bring problems such as racial biases. In all, the Bayes theorem has contributed a lot to the predictive policing algorithms and helped in preventing possible crimes in the future, with other benefits like strengthening police responsiveness and public confidence in the police.