A Framework for Exploring Crime Patterns and Insights in Quezon City: Analyzing Philippine News Articles Using Machine Learning Techniques

Philippines ranked
${25^{{\rm{th\;}}}}$
out of 193 countries in terms of criminality in the Global Organized Crime Index in 2023. Notably, in 2018, Quezon City alone reported 12,253 criminal cases, yet a disproportionately low number of cases were solved. It is evident that crimes harms individuals and its community and it is a significant threat to everyone. Using machine learning techniques, this study will train and test multiple clustering algorithms, including K-Means, Hierarchical, and DBSCAN clustering, as well as classification algorithms, including SVM, Naïve Bayes, and Decision Tree, to determine which algorithm is most effective in identifying crime trends and insights in Quezon City. The Silhouette Score will be utilized in this study to assess the models' performance for clustering methods, while evaluation metrics such as accuracy, precision, recall, and F-1 score will be used for classification algorithms. Time and space complexity will also be examined for classification algorithms. The dataset used in this study is scraped from GMA News. This study will present a framework for evaluating Philippine news articles in order to identify crime patterns and insights in Quezon City addressing the crime problem to help the LGU mitigate solution.