Crime Scene Prediction Using the Integration of K-Means Clustering and Support Vector Machine

Abstract

An increasing crime rate among urban residents has become a major concern over the last decade. Prevention can be done with advanced prediction based on criminal activities and locations. A publicly available dataset from kaggle.com was used in this research, consisting of 500 records of information, such as the coordinates of the crime locations and the types of crimes. An unsupervised machine learning algorithm, K-Means Clustering, is applied to group the data based on the locations of the reported crimes. Then, Support Vector Machine, a supervised machine learning algorithm, is applied to predict the potential crime locations. Thus, law enforcement agencies can make strategic plans and deploy their units to the predicted crime scenes, decreasing the chances of crimes being committed. Even though the integration of K-Means Clustering and Support Vector Machine for crime scene prediction only shows 0.65 accuracies, improvements can still be made for future work with larger datasets and integrating other machine learning algorithms.