<![CDATA[Reviews and ratings on apps store show how users perceive an app. This is an important aspect that companies must pay attention to in order to plan improvement strategies. However, there is an inconsistency between reviews and ratings, which makes it difficult to take corrective action. Detikcom, one of the main players in the digital media industry in Indonesia, also faces this similar problem. On Google Play Store platform, it is known that Detikcom's 1-star rating (13%) is one of the highest compared to its competitors. However, the inconsistency between ratings and reviews can be found frequently in the review section. Reflecting on this case, this study focuses on building a model that can identify sentiment using the Naïve Bayes Classifier method and identifying the main driving factors of each sentiment category using K-means Clustering and word cloud. Based on the results of the developed model, the tendency of Detikcom user sentiment is generally positive (67.16%) with a test accuracy of 87.84%. The presence of positive sentiment is based on keywords such as “accurate”, “trusty”, and “up to date”, while negative sentiment is based on keyword “ads”.]]>
Copyrights © 2025
