<![CDATA[The Digital Population Identity Application (IKD) is a government digital solution designed to facilitate public access to population data via mobile devices. This study aims to analyze the sentiment of user reviews for the Digital Population Identity (IKD) application available on the Google Play Store, employing the K-Nearest Neighbor (KNN) algorithm. The methodology encompasses data collection, preprocessing, division of data into training and testing sets, and the application of KNN with varying values of k. The findings reveal that the model utilizing k=17 with a 70:30 ratio attained an accuracy of 82%, alongside precision of 79%, recall of 82%, F1 Score of 79%, and a micro AUC value of 0.91. This model demonstrates effective identification of positive and negative sentiments; however, it encounters challenges in classifying neutral reviews. The study concludes that the Digital Population Identity (IKD) application is still regarded as relatively ineffective, given the significant percentage of negative sentiment, which accounts for 71.86% of the reviews. While the KNN algorithm proves effective for classification, enhancements in performance concerning the neutral class remain essential. Furthermore, variations in the value of k impact classification accuracy, and adjustments to the k parameter can enhance the model's overall performance.]]>
Copyrights © 2024
