Enhanced MacQueen's Algorithm for Identifying Diverse Crime Patterns in the City of Manila

MacQueen’s algorithm is a variant of the k-means algorithm used to determine clusters. However, the algorithm has its limitations that impact its accuracy and efficiency, resulting in suboptimal clustering. This study aimed to enhance MacQueen’s algorithm for analyzing diverse crime patterns in the city of Manila by addressing these limitations using Isolation Forest for outliers, Adaptive K-Means++ for algorithm initialization, and Gap Statistics to determine the optimal number of clusters. Isolation Forest was employed to detect and remove outliers from the dataset, as they significantly impact clustering results. Adaptive K-Means++ improved the initialization process by optimizing the placement of initial centroids, reducing the sensitivity of the algorithm to poor starting conditions. Gap Statistics was utilized to determine the optimal number of clusters, greatly enhancing the algorithm’s accuracy. The enhanced MacQueen’s algorithm demonstrated a significant overall improvement in clustering performance, resulting in more accurate and distinct clusters. The proposed enhancements effectively addressed the limitations of the traditional MacQueen’s algorithm, improving its accuracy and efficiency. This makes the enhanced algorithm highly applicable to real-world problems involving clustering.