Providing Real-time Crime Statistics in Indonesia Using Data Mining Approach

Abstract

Near-repeat crime effect is a condition where criminal activity tends to repeat in the near place and time. Over the past decades, most of the researchers performed the study on a different approach in crime forecasting. Using the crime data from the official provider, they use machine learning techniques to predict the crime pattern. In Indonesia, official crime data only available to the public annually. This makes it not easy to do crime forecasting. Thus, this research focuses on providing real-time crime data of criminal activity in Indonesia based on news available online using data mining techniques. News from various categories will be classified into a specific crime category. This research shows that data mining approach can support in providing real-time crime data based on online news as the dataset. As well, by implementing different types of classification techniques, this research able to suggest the most accurate method of classifying crime news using data mining.