Identifying Sentiment in User Reviews of Get Contact Application using Natural Language Processing

Analyzing user feedback on application updates may introduce a time-consuming task for the developers, since the task usually done manually. On the other hand, when users adjusting to these updates, it may elicit a range of feedbacks, both in favor and against the updates that have been implemented. Analyzing unfavorable reviews can be strategically utilized to enhance the app. Therefore, it is necessary to employ a straightforward, efficient, and accurate approach in analyzing user feedback. A potential approach is sentiment analysis, which can produce pertinent data by conducting multiple algorithmic tests. This study demonstrates the application of Natural Language Processing (NLP) in identifying sentiment in user feedbacks of Get Contact on Playstore. NLP is chosen for its remarkable proficiency in Sentiment analysis. The study involves multiple sequential steps: data collection, labeling, pre-processing, TF-IDF analysis, algorithm testing, result discussion, and summary. The examined algorithms include Naive Bayes, Deep Learning, Random Forest, K-Nearest Neighbor, and Decision Tree. Precision is chosen as the primary metric for evaluation, since it allows for more accurate predictions on bad reviews. The results indicate that the Naive Bayes approach fared better than the other algorithms in terms of precision and accuracy.