Based on 5853 comments across 5 topics, the analysis is as follows:

---

---

---

---

---

---

Based on the sentiment analysis and topic modeling performed on user reviews from the Play Store, we can draw several conclusions.

The sentiment classifier achieved an accuracy of 90.95%, indicating that it performs well in distinguishing between different sentiment categories. Based on the pie chart that shown in the sentiment menu, neutral sentiment is the most dominant (36.7%), followed by negative (32.5%) and positive (30.8%) sentiments.

The most frequently discussed topic is Topic 5, which includes keywords like . This suggests a high user interest or concern regarding this subject.

In terms of topic themes, the analysis indicates that users mostly talk about:

Technical issues (e.g., bugs, server problems), Positive experiences (e.g., love, enjoyment, creativity), New features or updates

Overall, the combination of sentiment classification and topic modeling offers a comprehensive understanding of how users perceive and discuss the latest updates or features in the application.

The most discussed topic revolves around technical issues, bugs, and server problems, with keywords such as "cant", "fix", "problem", and "server" frequently appearing. This indicates a clear concern from users about the game’s stability and functionality.

Although the majority of user comments are neutral (36.7%), the proportion of negative sentiment (32.5%) is notable and suggests growing dissatisfaction. Immediate action to address these issues can help prevent further negative sentiment.

It is recommended to prioritize bug fixes and server improvements to address the core concerns. Additionally, improving communication with users about ongoing fixes can help shift perceptions from neutral or negative toward a more positive sentiment.

Further model improvement and testing on new data are also encouraged to ensure robustness and enhance insight accuracy.