WEB-BASED ANALYSIS OF PUBLIC COMPLAINTS IN PERTUMBUKAN VILLAGE USING COSINE SIMILARITY AND TF-IDF TECHNIQUES

Abstract

This research focuses on the development of a web-based system for analyzing public complaints in Pertumbuhan Village using Cosine Similarity and Term Frequency-Inverse Document Frequency (TF-IDF) techniques. The system is intended to assist local authorities in efficiently managing and categorizing public complaints by identifying patterns and trends in the data. TF-IDF is applied to assign weights to words based on their frequency within a complaint, while Cosine Similarity is used to determine the similarity between different complaint texts. By integrating these methods, the system groups complaints with similar themes, enabling the local government to prioritize and address the most pressing issues quickly. The web-based platform also offers real-time data analysis, enabling more informed decision-making processes for resolving public grievances. The evaluation results show that this system can significantly reduce manual sorting tasks, improve complaint classification accuracy, and provide scalable solutions for handling large datasets. This approach facilitates a more transparent, efficient, and responsive public service delivery.