https://github.com/zeinhasan/sentiment-analysis-pedulilindungi-app-review
NLP & Text Classification using Deep Leaning and Machine Leaning
https://github.com/zeinhasan/sentiment-analysis-pedulilindungi-app-review
Last synced: 3 months ago
JSON representation
NLP & Text Classification using Deep Leaning and Machine Leaning
- Host: GitHub
- URL: https://github.com/zeinhasan/sentiment-analysis-pedulilindungi-app-review
- Owner: zeinhasan
- License: mit
- Created: 2024-08-06T07:33:24.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2024-08-06T07:35:27.000Z (10 months ago)
- Last Synced: 2025-01-02T11:28:41.489Z (5 months ago)
- Language: Jupyter Notebook
- Size: 1.51 MB
- Stars: 0
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Sentiment-Analysis-PeduliLindungi-App-Review
NLP & Text Classification using Deep Leaning and Machine Leaning
# Summary
Since early 2020, the COVID-19 pandemic has affected Indonesia. As the number of COVID-19 cases in the country increased, the government has taken strategic measures to manage the situation. One such measure was the implementation of mobility restrictions, known as “Lockdown,” to reduce virus transmission. However, the rapid spread of COVID-19 overwhelmed the government's monitoring efforts. In response, the Indonesian government launched an application called PeduliLindungi to track the spread of COVID-19. This application became a standard tool for tracking and minimizing the risk of transmission and is mandatory for Indonesian citizens.
PeduliLindungi has received both positive and negative feedback from users. This study aims to analyze user sentiment from reviews on the Google Play Store to gauge and improve the application's service quality. Many reviews, however, do not align with the given ratings, necessitating classification of reviews as either positive or negative.
The study employs algorithms such as Multilayer Perceptron, Support Vector Machine (SVM), and Naïve Bayes for classification. The Support Vector Machine algorithm achieved an accuracy of 95%, Naïve Bayes also achieved 95%, and Multilayer Perceptron achieved an accuracy of 96%. This indicates that the Multilayer Perceptron algorithm is the most effective of the three. The model developed can classify reviews very accurately, providing valuable insights for developers to understand user feedback and enhance the quality of PeduliLindungi.